diff --git "a/wandb/run-20240514_134421-rr61r5sa/files/output.log" "b/wandb/run-20240514_134421-rr61r5sa/files/output.log"
deleted file mode 100644--- "a/wandb/run-20240514_134421-rr61r5sa/files/output.log"
+++ /dev/null
@@ -1,21744 +0,0 @@
-WARNING ⚠️ no model scale passed. Assuming scale='x'.
-Transferred 1241/1241 items from pretrained weights
-train: WARNING ⚠️ /kaggle/working/datasets/VisDrone/VisDrone2019-DET-train/images/0000137_02220_d_0000163.jpg: 1 duplicate labels removed
-train: WARNING ⚠️ /kaggle/working/datasets/VisDrone/VisDrone2019-DET-train/images/0000140_00118_d_0000002.jpg: 1 duplicate labels removed
-train: WARNING ⚠️ /kaggle/working/datasets/VisDrone/VisDrone2019-DET-train/images/9999945_00000_d_0000114.jpg: 1 duplicate labels removed
-train: WARNING ⚠️ /kaggle/working/datasets/VisDrone/VisDrone2019-DET-train/images/9999987_00000_d_0000049.jpg: 1 duplicate labels removed
-train: Scanning /kaggle/working/datasets/VisDrone/VisDrone2019-DET-train/labels.cache... 6471 images, 0 backgrounds, 0 corrupt: 100%|██████████| 6471/6471 [00:00<?, ?it/s]
-val: Scanning /kaggle/working/datasets/VisDrone/VisDrone2019-DET-val/labels.cache... 548 images, 0 backgrounds, 0 corrupt: 100%|██████████| 548/548 [00:00<?, ?it/s]
-albumentations: Blur(p=0.01, blur_limit=(3, 7)), MedianBlur(p=0.01, blur_limit=(3, 7)), ToGray(p=0.01), CLAHE(p=0.01, clip_limit=(1, 4.0), tile_grid_size=(8, 8))
-Plotting labels to runs/detect/RTDETR/labels.jpg...
-optimizer: 'optimizer=auto' found, ignoring 'lr0=0.01' and 'momentum=0.937' and determining best 'optimizer', 'lr0' and 'momentum' automatically...
-optimizer: SGD(lr=0.01, momentum=0.9) with parameter groups 193 weight(decay=0.0), 256 weight(decay=0.0005), 276 bias(decay=0.0)
-Resuming training /kaggle/working/runs/detect/RTDETR/weights/last.pt from epoch 45 to 300 total epochs
-  0%|          | 0/809 [00:00<?, ?it/s]
-TensorBoard: model graph visualization added ✅
-Image sizes 640 train, 640 val
-Using 4 dataloader workers
-Logging results to runs/detect/RTDETR
-Starting training for 300 epochs...
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     45/300      14.9G       1.22     0.4328     0.1445        445        640: 100%|██████████| 809/809 [15:48<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  99%|█████████▊| 68/69 [00:18<00:00,  4.19it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:18<00:00,  3.70it/s]
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     46/300      13.4G      1.234     0.4189     0.1434        158        640: 100%|██████████| 809/809 [15:51<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.05it/s]
-                   all        548      38759      0.503      0.326      0.333      0.178
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     47/300      13.8G      1.199      0.431     0.1379        203        640: 100%|██████████| 809/809 [15:49<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.04it/s]
-                   all        548      38759       0.49      0.323      0.327      0.173
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-     48/300      14.2G      1.686     0.3222     0.3328        406        640:   0%|          | 1/809 [00:01<15:43,  1.17s/it]
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     48/300      14.3G      1.195     0.4312     0.1378        145        640: 100%|██████████| 809/809 [15:51<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  93%|█████████▎| 64/69 [00:16<00:01,  4.10it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.03it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     49/300      13.5G      1.202     0.4336      0.139        446        640: 100%|██████████| 809/809 [15:49<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:16<00:00,  4.12it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:16<00:00,  4.07it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     50/300      13.9G      1.173     0.4426     0.1369        265        640: 100%|██████████| 809/809 [15:51<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  99%|█████████▊| 68/69 [00:17<00:00,  4.17it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.02it/s]
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     51/300      13.7G      1.147     0.4495     0.1332        185        640: 100%|██████████| 809/809 [15:50<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.01it/s]
-                   all        548      38759      0.492      0.336      0.343      0.184
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     52/300      13.8G      1.178     0.4365     0.1347        102        640: 100%|██████████| 809/809 [15:50<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.06it/s]
-                   all        548      38759      0.471      0.327      0.327      0.175
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     53/300      14.1G       1.18     0.4427     0.1356        277        640: 100%|██████████| 809/809 [15:49<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:16<00:00,  3.96it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.05it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     54/300      13.6G      1.166     0.4468     0.1341        180        640: 100%|██████████| 809/809 [15:47<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:16<00:00,  4.08it/s]
-                   all        548      38759      0.495      0.339      0.341      0.192
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     55/300      14.2G      1.175     0.4348     0.1289        131        640: 100%|██████████| 809/809 [15:49<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:16<00:00,  4.07it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.06it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     56/300      13.5G      1.162     0.4464     0.1344        375        640: 100%|██████████| 809/809 [15:50<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.05it/s]
-                   all        548      38759      0.472      0.335      0.331      0.173
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     57/300      14.1G      1.121     0.4535     0.1226        303        640: 100%|██████████| 809/809 [15:49<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.06it/s]
-                   all        548      38759       0.52      0.342      0.352      0.192
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     58/300      14.6G       1.17     0.4392     0.1349        359        640: 100%|██████████| 809/809 [15:51<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  90%|████████▉ | 62/69 [00:15<00:01,  4.06it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|████████���█| 69/69 [00:17<00:00,  4.05it/s]
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     59/300      13.5G      1.199     0.4317     0.1348        280        640: 100%|██████████| 809/809 [15:51<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  93%|█████████▎| 64/69 [00:16<00:01,  3.99it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.02it/s]
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     60/300      13.9G       1.14     0.4497      0.125        252        640: 100%|██████████| 809/809 [15:52<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  99%|█████████▊| 68/69 [00:16<00:00,  4.09it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:16<00:00,  4.06it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     61/300      14.1G      1.158     0.4468     0.1291        307        640: 100%|██████████| 809/809 [15:51<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.03it/s]
-                   all        548      38759      0.477      0.328       0.33      0.176
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     62/300      13.7G      1.135     0.4512     0.1215        472        640: 100%|██████████| 809/809 [15:50<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:00,  4.09it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.06it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     63/300      13.5G      1.176     0.4411     0.1338        252        640: 100%|██████████| 809/809 [15:51<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.04it/s]
-                   all        548      38759      0.506      0.351      0.354      0.197
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     64/300      13.5G      1.183     0.4341     0.1327        157        640: 100%|██████████| 809/809 [15:51<00:00,  1.18s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:00,  4.03it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.04it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     65/300      13.9G       1.14     0.4542     0.1257        284        640: 100%|██████████| 809/809 [15:50<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.01it/s]
-                   all        548      38759      0.516      0.355      0.361      0.198
-  0%|          | 0/809 [00:00<?, ?it/s]
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     66/300      13.7G      1.135     0.4484     0.1244        336        640: 100%|██████████| 809/809 [15:49<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:16<00:00,  4.02it/s]
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:16<00:00,  4.06it/s]
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     67/300      13.9G       1.09     0.4677     0.1178        288        640: 100%|██████████| 809/809 [15:48<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:16<00:00,  4.07it/s]
-                   all        548      38759      0.532      0.359      0.368      0.208
-      Epoch    GPU_mem  giou_loss   cls_loss    l1_loss  Instances       Size
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-     68/300      13.9G      1.127      0.448     0.1217        161        640: 100%|██████████| 809/809 [15:49<00:00,  1.17s/it]
-
-
-
-
-
-
-
-
-
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.05it/s]
-                   all        548      38759      0.476      0.334      0.329      0.161
-  0%|          | 0/809 [00:00<?, ?it/s]/opt/conda/lib/python3.10/site-packages/torch/autograd/__init__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward pass
-     69/300      13.9G      1.377     0.3365     0.1802        481        640:   0%|          | 1/809 [00:01<16:51,  1.25s/it]
-     69/300      13.9G      1.321     0.3837     0.1463        361        640:   0%|          | 3/809 [00:03<15:47,  1.18s/it]
-     69/300      13.9G      1.353     0.3713     0.1428        197        640:   0%|          | 4/809 [00:04<16:16,  1.21s/it]
-     69/300      13.9G      1.241      0.415     0.1323        292        640:   1%|          | 6/809 [00:07<15:50,  1.18s/it]
-     69/300      13.9G      1.221     0.4179     0.1314        131        640:   1%|          | 8/809 [00:09<15:38,  1.17s/it]
-     69/300      13.9G      1.193      0.424     0.1258        164        640:   1%|          | 9/809 [00:10<15:41,  1.18s/it]
-     69/300      13.9G      1.167     0.4339     0.1238        133        640:   1%|▏         | 11/809 [00:12<15:23,  1.16s/it]
-     69/300      13.9G      1.186     0.4272      0.124        250        640:   2%|▏         | 13/809 [00:15<15:30,  1.17s/it]
-     69/300      13.9G      1.178      0.431     0.1231        169        640:   2%|▏         | 15/809 [00:17<15:32,  1.17s/it]
-     69/300      13.9G      1.186     0.4302     0.1235        259        640:   2%|▏         | 16/809 [00:18<15:38,  1.18s/it]
-     69/300      13.9G        1.2     0.4251     0.1243        424        640:   2%|▏         | 18/809 [00:21<15:24,  1.17s/it]
-     69/300      13.9G      1.225     0.4152      0.125        277        640:   2%|▏         | 20/809 [00:23<15:16,  1.16s/it]
-     69/300      13.9G      1.246     0.4079      0.125        190        640:   3%|▎         | 21/809 [00:24<15:25,  1.17s/it]
-     69/300      13.9G      1.258     0.4062     0.1262        149        640:   3%|▎         | 23/809 [00:27<15:20,  1.17s/it]
-     69/300      13.9G      1.269     0.4021     0.1256        203        640:   3%|▎         | 25/809 [00:29<15:45,  1.21s/it]
-     69/300      13.9G      1.266     0.4029     0.1249        277        640:   3%|▎         | 26/809 [00:30<15:30,  1.19s/it]
-     69/300      13.9G      1.266     0.4026     0.1235        210        640:   3%|▎         | 28/809 [00:33<15:32,  1.19s/it]
-     69/300      13.9G      1.259     0.4047     0.1218        152        640:   4%|▎         | 30/809 [00:35<15:06,  1.16s/it]
-     69/300      13.9G       1.27     0.4016     0.1227        249        640:   4%|▍         | 32/809 [00:37<15:03,  1.16s/it]
-     69/300      13.9G       1.26     0.4042     0.1222         64        640:   4%|▍         | 33/809 [00:38<15:09,  1.17s/it]
-     69/300      13.9G      1.253     0.4054     0.1206         82        640:   4%|▍         | 35/809 [00:41<15:24,  1.19s/it]
-     69/300      13.9G      1.253     0.4053     0.1263        198        640:   5%|▍         | 37/809 [00:43<15:12,  1.18s/it]
-     69/300      13.9G      1.251     0.4051     0.1256        233        640:   5%|▍         | 38/809 [00:44<15:11,  1.18s/it]
-     69/300      13.9G      1.248     0.4053     0.1243        397        640:   5%|▍         | 40/809 [00:47<14:58,  1.17s/it]
-     69/300      13.9G      1.239     0.4082      0.123        149        640:   5%|▌         | 42/809 [00:49<14:53,  1.16s/it]
-     69/300      13.9G       1.24     0.4071     0.1227        283        640:   5%|▌         | 43/809 [00:50<14:47,  1.16s/it]
-     69/300      13.9G      1.221     0.4127     0.1206        107        640:   6%|▌         | 45/809 [00:52<14:40,  1.15s/it]
-     69/300      13.9G      1.218      0.415     0.1197        181        640:   6%|▌         | 47/809 [00:55<14:35,  1.15s/it]
-     69/300      13.9G      1.216     0.4154     0.1183        128        640:   6%|▌         | 49/809 [00:57<15:09,  1.20s/it]
-     69/300      13.9G      1.214     0.4168     0.1184        191        640:   6%|▌         | 50/809 [00:58<14:54,  1.18s/it]
-     69/300      13.9G      1.208     0.4185     0.1181        166        640:   6%|▋         | 52/809 [01:01<14:52,  1.18s/it]
-     69/300      13.9G      1.204     0.4185     0.1174        351        640:   7%|▋         | 54/809 [01:03<14:34,  1.16s/it]
-     69/300      13.9G      1.208     0.4167     0.1169        382        640:   7%|▋         | 56/809 [01:05<14:23,  1.15s/it]
-     69/300      13.9G      1.202     0.4178     0.1164        192        640:   7%|▋         | 57/809 [01:06<14:30,  1.16s/it]
-     69/300      13.9G        1.2      0.418     0.1167        273        640:   7%|▋         | 59/809 [01:09<14:33,  1.17s/it]
-     69/300      13.9G      1.198      0.419     0.1166         61        640:   8%|▊         | 61/809 [01:11<14:34,  1.17s/it]
-     69/300      13.9G      1.197     0.4189     0.1166        203        640:   8%|▊         | 62/809 [01:12<14:24,  1.16s/it]
-     69/300      13.9G      1.194     0.4193     0.1161        155        640:   8%|▊         | 64/809 [01:15<14:27,  1.16s/it]
-     69/300      13.9G      1.188     0.4215     0.1161        254        640:   8%|▊         | 66/809 [01:17<14:37,  1.18s/it]
-     69/300      13.9G       1.18     0.4245     0.1153        150        640:   8%|▊         | 68/809 [01:19<14:21,  1.16s/it]
-     69/300      13.9G      1.178     0.4247     0.1154        144        640:   9%|▊         | 69/809 [01:20<14:27,  1.17s/it]
-     69/300      13.9G      1.169     0.4282     0.1145        120        640:   9%|▉         | 71/809 [01:23<14:17,  1.16s/it]
-     69/300      13.9G      1.169     0.4289     0.1147        142        640:   9%|▉         | 73/809 [01:25<14:30,  1.18s/it]
-     69/300      13.9G       1.17     0.4287     0.1149        310        640:   9%|▉         | 74/809 [01:26<14:18,  1.17s/it]
-     69/300      13.9G      1.173     0.4281     0.1152        294        640:   9%|▉         | 76/809 [01:29<14:12,  1.16s/it]
-     69/300      13.9G       1.17     0.4299     0.1151        280        640:  10%|▉         | 78/809 [01:31<14:08,  1.16s/it]
-     69/300      13.9G      1.167     0.4312     0.1148        204        640:  10%|▉         | 80/809 [01:33<14:12,  1.17s/it]
-     69/300      13.9G      1.165     0.4313     0.1147        193        640:  10%|█         | 81/809 [01:35<14:18,  1.18s/it]
-     69/300      13.9G      1.166     0.4317      0.115        445        640:  10%|█         | 83/809 [01:37<14:11,  1.17s/it]
-     69/300      13.9G      1.163     0.4332     0.1146        121        640:  11%|█         | 85/809 [01:39<14:04,  1.17s/it]
-     69/300      13.9G      1.173      0.431     0.1187        786        640:  11%|█         | 86/809 [01:40<14:16,  1.18s/it]
-     69/300      13.9G      1.179     0.4291     0.1196        381        640:  11%|█         | 88/809 [01:43<13:58,  1.16s/it]
-     69/300      13.9G       1.18     0.4291     0.1197        522        640:  11%|█         | 90/809 [01:45<13:56,  1.16s/it]
-     69/300      13.9G       1.18     0.4285     0.1199        197        640:  11%|█▏        | 92/809 [01:47<13:44,  1.15s/it]
-     69/300      13.9G      1.178      0.429     0.1196        254        640:  11%|█▏        | 93/809 [01:48<13:44,  1.15s/it]
-     69/300      13.9G      1.174     0.4311     0.1191        291        640:  12%|█▏        | 95/809 [01:51<13:56,  1.17s/it]
-     69/300      13.9G      1.169     0.4318     0.1184        201        640:  12%|█▏        | 97/809 [01:53<13:53,  1.17s/it]
-     69/300      13.9G      1.169     0.4314     0.1183        105        640:  12%|█▏        | 98/809 [01:54<13:44,  1.16s/it]
-     69/300      13.9G      1.173     0.4297     0.1196        255        640:  12%|█▏        | 100/809 [01:57<13:58,  1.18s/it]
-     69/300      13.9G      1.178     0.4288     0.1203        351        640:  13%|█▎        | 102/809 [01:59<13:45,  1.17s/it]
-     69/300      13.9G      1.184     0.4271     0.1228        750        640:  13%|█▎        | 103/809 [02:00<14:09,  1.20s/it]
-     69/300      13.9G      1.184     0.4279     0.1227        458        640:  13%|█▎        | 105/809 [02:03<14:00,  1.19s/it]
-     69/300      13.9G      1.181     0.4288     0.1227        126        640:  13%|█▎        | 107/809 [02:05<13:46,  1.18s/it]
-     69/300      13.9G      1.181     0.4284     0.1228        174        640:  13%|█▎        | 109/809 [02:07<13:41,  1.17s/it]
-     69/300      13.9G       1.18     0.4286     0.1226        265        640:  14%|█▎        | 110/809 [02:09<13:48,  1.19s/it]
-     69/300      13.9G      1.175       0.43      0.122        195        640:  14%|█▍        | 112/809 [02:11<13:31,  1.16s/it]
-     69/300      13.9G      1.174     0.4302     0.1215        255        640:  14%|█▍        | 114/809 [02:13<13:44,  1.19s/it]
-     69/300      13.9G      1.177     0.4292     0.1225        423        640:  14%|█▍        | 115/809 [02:14<13:37,  1.18s/it]
-     69/300      13.9G      1.175     0.4296     0.1227        412        640:  14%|█▍        | 117/809 [02:17<13:20,  1.16s/it]
-     69/300      13.9G      1.175     0.4293     0.1225        307        640:  15%|█▍        | 119/809 [02:19<13:25,  1.17s/it]
-     69/300      13.9G      1.175     0.4295     0.1226        152        640:  15%|█▍        | 121/809 [02:21<13:21,  1.16s/it]
-     69/300      13.9G      1.173     0.4302     0.1222        184        640:  15%|█▌        | 122/809 [02:23<13:22,  1.17s/it]
-     69/300      13.9G      1.171     0.4303     0.1217        242        640:  15%|█▌        | 124/809 [02:25<13:05,  1.15s/it]
-     69/300      13.9G      1.167     0.4312     0.1212        214        640:  16%|█▌        | 126/809 [02:27<12:58,  1.14s/it]
-     69/300      13.9G      1.167     0.4308     0.1214        331        640:  16%|█▌        | 128/809 [02:29<13:06,  1.15s/it]
-     69/300      13.9G      1.166     0.4309     0.1212        290        640:  16%|█▌        | 129/809 [02:31<13:08,  1.16s/it]
-     69/300      13.9G      1.166     0.4311     0.1211        263        640:  16%|█▌        | 131/809 [02:33<13:02,  1.15s/it]
-     69/300      13.9G      1.166     0.4313     0.1208        199        640:  16%|█▋        | 133/809 [02:35<13:01,  1.16s/it]
-     69/300      13.9G      1.166     0.4311     0.1207        331        640:  17%|█▋        | 135/809 [02:38<13:05,  1.17s/it]
-     69/300      13.9G      1.166     0.4307     0.1207        344        640:  17%|█▋        | 136/809 [02:39<13:06,  1.17s/it]
-     69/300      13.9G      1.172      0.429     0.1223        393        640:  17%|█▋        | 138/809 [02:41<13:13,  1.18s/it]
-     69/300      13.9G      1.175      0.428     0.1228        444        640:  17%|█▋        | 140/809 [02:43<12:57,  1.16s/it]
-     69/300      13.9G      1.175     0.4282     0.1227        125        640:  17%|█▋        | 141/809 [02:45<12:55,  1.16s/it]
-     69/300      13.9G      1.176     0.4278     0.1225        220        640:  18%|█▊        | 143/809 [02:47<12:46,  1.15s/it]
-     69/300      13.9G      1.175     0.4278     0.1225        182        640:  18%|█▊        | 145/809 [02:49<12:53,  1.16s/it]
-     69/300      13.9G      1.174     0.4284     0.1223        247        640:  18%|█▊        | 147/809 [02:52<12:56,  1.17s/it]
-     69/300      13.9G      1.174     0.4284     0.1222        251        640:  18%|█▊        | 148/809 [02:53<12:51,  1.17s/it]
-     69/300      13.9G      1.171     0.4296     0.1221        411        640:  19%|█▊        | 150/809 [02:55<12:54,  1.18s/it]
-     69/300      13.9G      1.171     0.4298     0.1219        299        640:  19%|█▉        | 152/809 [02:57<12:43,  1.16s/it]
-     69/300      13.9G       1.17     0.4304     0.1218        276        640:  19%|█▉        | 153/809 [02:59<13:26,  1.23s/it]
-     69/300      13.9G      1.168     0.4308     0.1213        270        640:  19%|█▉        | 155/809 [03:01<13:04,  1.20s/it]
-     69/300      13.9G      1.167     0.4316     0.1211        135        640:  19%|█▉        | 157/809 [03:03<12:47,  1.18s/it]
-     69/300      13.9G      1.165      0.432     0.1209        310        640:  20%|█▉        | 158/809 [03:05<12:38,  1.17s/it]
-     69/300      13.9G      1.167     0.4313     0.1214        252        640:  20%|█▉        | 160/809 [03:07<12:47,  1.18s/it]
-     69/300      13.9G      1.164     0.4321     0.1213        249        640:  20%|██        | 162/809 [03:09<12:55,  1.20s/it]
-     69/300      13.9G      1.164     0.4323     0.1211        240        640:  20%|██        | 163/809 [03:11<12:45,  1.19s/it]
-     69/300      13.9G      1.168     0.4312     0.1227        406        640:  20%|██        | 165/809 [03:13<12:33,  1.17s/it]
-     69/300      13.9G       1.17      0.431     0.1227        280        640:  21%|██        | 167/809 [03:15<12:25,  1.16s/it]
-     69/300      13.9G      1.173     0.4307     0.1266        468        640:  21%|██        | 169/809 [03:18<12:58,  1.22s/it]
-     69/300      13.9G      1.176       0.43     0.1271        501        640:  21%|██        | 170/809 [03:19<12:42,  1.19s/it]
-     69/300      13.9G      1.178     0.4295      0.127        302        640:  21%|██▏       | 172/809 [03:21<12:21,  1.16s/it]
-     69/300      13.9G      1.182     0.4288     0.1293        271        640:  22%|██▏       | 174/809 [03:23<12:19,  1.16s/it]
-     69/300      13.9G      1.182     0.4291     0.1291        236        640:  22%|██▏       | 175/809 [03:25<12:13,  1.16s/it]
-     69/300      13.9G      1.187     0.4279     0.1308        372        640:  22%|██▏       | 177/809 [03:27<12:25,  1.18s/it]
-     69/300      13.9G      1.188     0.4277     0.1314        207        640:  22%|██▏       | 179/809 [03:29<12:13,  1.17s/it]
-     69/300      13.9G      1.187      0.428     0.1313        145        640:  22%|██▏       | 181/809 [03:32<12:05,  1.16s/it]
-     69/300      13.9G      1.186     0.4284      0.131        208        640:  22%|██▏       | 182/809 [03:33<12:00,  1.15s/it]
-     69/300      13.9G      1.186     0.4293     0.1314        126        640:  23%|██▎       | 184/809 [03:35<11:57,  1.15s/it]
-     69/300      13.9G      1.184     0.4299     0.1311        301        640:  23%|██▎       | 186/809 [03:37<11:56,  1.15s/it]
-     69/300      13.9G      1.181     0.4306     0.1306        208        640:  23%|██▎       | 188/809 [03:40<12:05,  1.17s/it]
-     69/300      13.9G      1.179     0.4309     0.1304        207        640:  23%|██▎       | 189/809 [03:41<12:12,  1.18s/it]
-     69/300      13.9G      1.178      0.431     0.1303        432        640:  24%|██▎       | 191/809 [03:43<11:58,  1.16s/it]
-     69/300      13.9G      1.179     0.4306     0.1304        310        640:  24%|██▍       | 193/809 [03:45<11:56,  1.16s/it]
-     69/300      13.9G      1.181     0.4301     0.1307        163        640:  24%|██▍       | 194/809 [03:47<11:47,  1.15s/it]
-     69/300      13.9G      1.181     0.4299     0.1303        218        640:  24%|██▍       | 196/809 [03:49<11:48,  1.16s/it]
-     69/300      13.9G      1.184     0.4288     0.1308        228        640:  24%|██▍       | 198/809 [03:51<11:39,  1.15s/it]
-     69/300      13.9G      1.183     0.4288     0.1305        163        640:  25%|██▍       | 200/809 [03:53<11:34,  1.14s/it]
-     69/300      13.9G      1.184     0.4284     0.1306        173        640:  25%|██▍       | 201/809 [03:55<12:10,  1.20s/it]
-     69/300      13.9G      1.182     0.4297     0.1305        386        640:  25%|██▌       | 203/809 [03:57<11:53,  1.18s/it]
-     69/300      13.9G      1.182     0.4293     0.1304        377        640:  25%|██▌       | 205/809 [03:59<11:37,  1.15s/it]
-     69/300      13.9G      1.182     0.4294     0.1303        270        640:  26%|██▌       | 207/809 [04:02<11:41,  1.17s/it]
-     69/300      13.9G       1.18     0.4297     0.1301        194        640:  26%|██▌       | 208/809 [04:03<11:37,  1.16s/it]
-     69/300      13.9G       1.18     0.4301       0.13        190        640:  26%|██▌       | 210/809 [04:05<11:53,  1.19s/it]
-     69/300      13.9G      1.183     0.4295     0.1305        343        640:  26%|██▌       | 212/809 [04:08<11:48,  1.19s/it]
-     69/300      13.9G      1.184     0.4292     0.1306        134        640:  26%|██▋       | 213/809 [04:09<11:42,  1.18s/it]
-     69/300      13.9G      1.187     0.4282      0.131        115        640:  27%|██▋       | 215/809 [04:11<11:42,  1.18s/it]
-     69/300      13.9G      1.187     0.4286     0.1308        159        640:  27%|██▋       | 217/809 [04:14<11:40,  1.18s/it]
-     69/300      13.9G      1.188     0.4282     0.1309        341        640:  27%|██▋       | 218/809 [04:15<11:27,  1.16s/it]
-     69/300      13.9G       1.19     0.4275     0.1308        329        640:  27%|██▋       | 220/809 [04:17<11:25,  1.16s/it]
-     69/300      13.9G      1.189     0.4279     0.1307        128        640:  27%|██▋       | 222/809 [04:19<11:28,  1.17s/it]
-     69/300      13.9G      1.189     0.4277     0.1308        390        640:  28%|██▊       | 224/809 [04:22<11:23,  1.17s/it]
-     69/300      13.9G      1.189     0.4276     0.1307        163        640:  28%|██▊       | 225/809 [04:23<11:30,  1.18s/it]
-     69/300      13.9G      1.191     0.4268     0.1308        272        640:  28%|██▊       | 227/809 [04:25<11:14,  1.16s/it]
-     69/300      13.9G       1.19     0.4268     0.1306        213        640:  28%|██▊       | 229/809 [04:28<11:04,  1.14s/it]
-     69/300      13.9G      1.192     0.4261      0.131        646        640:  28%|██▊       | 230/809 [04:29<11:12,  1.16s/it]
-     69/300      13.9G      1.194     0.4252     0.1312        629        640:  29%|██▊       | 232/809 [04:31<11:14,  1.17s/it]
-     69/300      13.9G      1.196     0.4245     0.1312        328        640:  29%|██▉       | 234/809 [04:33<11:10,  1.17s/it]
-     69/300      13.9G      1.195     0.4247      0.131        225        640:  29%|██▉       | 236/809 [04:36<11:01,  1.15s/it]
-     69/300      13.9G      1.195     0.4245      0.131        279        640:  29%|██▉       | 237/809 [04:37<11:09,  1.17s/it]
-     69/300      13.9G      1.195     0.4244     0.1311        286        640:  30%|██▉       | 239/809 [04:39<11:00,  1.16s/it]
-     69/300      13.9G      1.196     0.4241     0.1312        157        640:  30%|██▉       | 241/809 [04:42<11:03,  1.17s/it]
-     69/300      13.9G      1.197     0.4234     0.1312        283        640:  30%|██▉       | 242/809 [04:44<11:04,  1.17s/it]
-     69/300      13.9G      1.198      0.423     0.1311        275        640:  30%|███       | 244/809 [04:45<10:57,  1.16s/it]
-     69/300      13.9G      1.197     0.4234     0.1309         86        640:  30%|███       | 246/809 [04:47<11:00,  1.17s/it]
-     69/300      13.9G      1.197     0.4235      0.131         99        640:  31%|███       | 248/809 [04:50<10:54,  1.17s/it]
-     69/300      13.9G      1.199      0.423     0.1313        387        640:  31%|███       | 249/809 [04:51<11:07,  1.19s/it]
-     69/300      13.9G      1.197     0.4235     0.1312        334        640:  31%|███       | 251/809 [04:53<11:04,  1.19s/it]
-     69/300      13.9G      1.197     0.4233     0.1313        375        640:  31%|███▏      | 253/809 [04:56<10:48,  1.17s/it]
-     69/300      13.9G      1.197     0.4235     0.1312        292        640:  31%|███▏      | 254/809 [04:57<10:46,  1.16s/it]
-     69/300      13.9G      1.195     0.4241     0.1311        155        640:  32%|███▏      | 256/809 [04:59<10:46,  1.17s/it]
-     69/300      13.9G      1.196     0.4241     0.1312        256        640:  32%|███▏      | 258/809 [05:02<10:52,  1.18s/it]
-     69/300      13.9G      1.198     0.4236     0.1315        189        640:  32%|███▏      | 260/809 [05:04<10:47,  1.18s/it]
-     69/300      13.9G      1.197     0.4238     0.1314        229        640:  32%|███▏      | 261/809 [05:05<10:38,  1.17s/it]
-     69/300      13.9G      1.199     0.4232     0.1318        379        640:  33%|███▎      | 263/809 [05:07<10:31,  1.16s/it]
-     69/300      13.9G      1.199     0.4231     0.1317        232        640:  33%|███▎      | 265/809 [05:10<10:32,  1.16s/it]
-     69/300      13.9G        1.2     0.4227     0.1316        185        640:  33%|███▎      | 266/809 [05:11<10:31,  1.16s/it]
-     69/300      13.9G        1.2     0.4227     0.1315        197        640:  33%|███▎      | 268/809 [05:13<10:32,  1.17s/it]
-     69/300      13.9G        1.2     0.4224     0.1314        149        640:  33%|███▎      | 270/809 [05:16<10:40,  1.19s/it]
-     69/300      13.9G      1.201     0.4222     0.1313         81        640:  34%|███▎      | 272/809 [05:18<10:26,  1.17s/it]
-     69/300      13.9G      1.201     0.4221     0.1312        387        640:  34%|███▎      | 273/809 [05:19<10:31,  1.18s/it]
-     69/300      13.9G        1.2     0.4225     0.1312        356        640:  34%|███▍      | 275/809 [05:21<10:18,  1.16s/it]
-     69/300      13.9G      1.198     0.4228     0.1308        289        640:  34%|███▍      | 277/809 [05:24<10:12,  1.15s/it]
-     69/300      13.9G      1.196     0.4238     0.1305        191        640:  34%|███▍      | 279/809 [05:26<10:11,  1.15s/it]
-     69/300      13.9G      1.195     0.4239     0.1304        204        640:  35%|███▍      | 280/809 [05:27<10:11,  1.16s/it]
-     69/300      13.9G      1.194     0.4244     0.1302        363        640:  35%|███▍      | 282/809 [05:29<10:11,  1.16s/it]
-     69/300      13.9G      1.194     0.4243      0.131         95        640:  35%|███▌      | 284/809 [05:32<10:02,  1.15s/it]
-     69/300      13.9G      1.193     0.4249     0.1307        151        640:  35%|███▌      | 286/809 [05:34<09:57,  1.14s/it]
-     69/300      13.9G      1.192     0.4253     0.1305        189        640:  35%|███▌      | 287/809 [05:35<10:06,  1.16s/it]
-     69/300      13.9G      1.191     0.4257     0.1305        414        640:  36%|███▌      | 289/809 [05:38<10:20,  1.19s/it]
-     69/300      13.9G      1.191     0.4257     0.1305        172        640:  36%|███▌      | 291/809 [05:40<10:11,  1.18s/it]
-     69/300      13.9G      1.191     0.4257     0.1304        122        640:  36%|███▌      | 292/809 [05:41<10:02,  1.17s/it]
-     69/300      13.9G      1.192     0.4255     0.1303        168        640:  36%|███▋      | 294/809 [05:44<10:21,  1.21s/it]
-     69/300      13.9G      1.192     0.4256     0.1303        239        640:  36%|███▋      | 295/809 [05:45<10:34,  1.23s/it]
-     69/300      13.9G      1.193     0.4253     0.1302        245        640:  37%|███▋      | 297/809 [05:48<10:39,  1.25s/it]
-     69/300      13.9G      1.193     0.4251     0.1301        396        640:  37%|███▋      | 299/809 [05:50<10:10,  1.20s/it]
-     69/300      13.9G      1.192     0.4251       0.13        360        640:  37%|███▋      | 300/809 [05:51<09:59,  1.18s/it]
-     69/300      13.9G      1.192     0.4253     0.1301        267        640:  37%|███▋      | 302/809 [05:53<10:13,  1.21s/it]
-     69/300      13.9G      1.192     0.4253       0.13        317        640:  38%|███▊      | 304/809 [05:56<09:50,  1.17s/it]
-     69/300      13.9G      1.192     0.4254     0.1303        342        640:  38%|███▊      | 306/809 [05:58<09:46,  1.17s/it]
-     69/300      13.9G      1.193     0.4252     0.1304        387        640:  38%|███▊      | 307/809 [05:59<09:42,  1.16s/it]
-     69/300      13.9G       1.19      0.426     0.1302        115        640:  38%|███▊      | 309/809 [06:01<09:35,  1.15s/it]
-     69/300      13.9G       1.19     0.4261     0.1301        193        640:  38%|███▊      | 311/809 [06:04<09:34,  1.15s/it]
-     69/300      13.9G       1.19     0.4261     0.1301        296        640:  39%|███▊      | 312/809 [06:05<09:36,  1.16s/it]
-     69/300      13.9G      1.188     0.4268     0.1299        276        640:  39%|███▉      | 314/809 [06:07<10:01,  1.21s/it]
-     69/300      13.9G      1.187     0.4271     0.1298        414        640:  39%|███▉      | 316/809 [06:10<09:39,  1.18s/it]
-     69/300      13.9G      1.187     0.4271     0.1297        250        640:  39%|███▉      | 318/809 [06:12<09:32,  1.17s/it]
-     69/300      13.9G      1.186     0.4274     0.1296        139        640:  39%|███▉      | 319/809 [06:13<09:26,  1.16s/it]
-     69/300      13.9G      1.185     0.4277     0.1294        181        640:  40%|███▉      | 321/809 [06:16<09:32,  1.17s/it]
-     69/300      13.9G      1.186     0.4273     0.1298        318        640:  40%|███▉      | 323/809 [06:18<09:32,  1.18s/it]
-     69/300      13.9G      1.186     0.4276     0.1297        276        640:  40%|████      | 324/809 [06:19<09:28,  1.17s/it]
-     69/300      13.9G      1.185     0.4277     0.1295        234        640:  40%|████      | 326/809 [06:21<09:30,  1.18s/it]
-     69/300      13.9G      1.186     0.4275     0.1298        285        640:  41%|████      | 328/809 [06:24<09:18,  1.16s/it]
-     69/300      13.9G      1.184     0.4283     0.1297        410        640:  41%|████      | 330/809 [06:26<09:17,  1.16s/it]
-     69/300      13.9G      1.184     0.4284     0.1296        237        640:  41%|████      | 331/809 [06:27<09:13,  1.16s/it]
-     69/300      13.9G      1.183     0.4285     0.1293        280        640:  41%|████      | 333/809 [06:30<09:04,  1.14s/it]
-     69/300      13.9G      1.182     0.4289     0.1293        311        640:  41%|████▏     | 335/809 [06:32<09:00,  1.14s/it]
-     69/300      13.9G      1.183     0.4287     0.1294        254        640:  42%|████▏     | 337/809 [06:34<09:08,  1.16s/it]
-     69/300      13.9G      1.184     0.4283     0.1294        218        640:  42%|████▏     | 338/809 [06:35<09:16,  1.18s/it]
-     69/300      13.9G      1.186     0.4276     0.1295        278        640:  42%|████▏     | 340/809 [06:38<09:13,  1.18s/it]
-     69/300      13.9G      1.187     0.4274     0.1296        199        640:  42%|████▏     | 342/809 [06:40<09:05,  1.17s/it]
-     69/300      13.9G      1.188     0.4271     0.1296        146        640:  42%|████▏     | 343/809 [06:41<09:05,  1.17s/it]
-     69/300      13.9G      1.188     0.4267     0.1298        121        640:  43%|████▎     | 345/809 [06:44<09:03,  1.17s/it]
-     69/300      13.9G      1.187     0.4271     0.1295        349        640:  43%|████▎     | 347/809 [06:46<08:53,  1.15s/it]
-     69/300      13.9G      1.185     0.4276     0.1293        321        640:  43%|████▎     | 349/809 [06:48<08:51,  1.16s/it]
-     69/300      13.9G      1.185     0.4274     0.1294        335        640:  43%|████▎     | 350/809 [06:49<08:58,  1.17s/it]
-     69/300      13.9G      1.184     0.4278     0.1292        200        640:  44%|████▎     | 352/809 [06:52<09:20,  1.23s/it]
-     69/300      13.9G      1.183     0.4282      0.129        118        640:  44%|████▎     | 353/809 [06:53<09:15,  1.22s/it]
-     69/300      13.9G      1.182     0.4286     0.1288        251        640:  44%|████▍     | 355/809 [06:56<09:11,  1.21s/it]
-     69/300      13.9G       1.18     0.4292     0.1285        222        640:  44%|████▍     | 357/809 [06:58<09:02,  1.20s/it]
-     69/300      13.9G      1.179     0.4294     0.1283        153        640:  44%|████▍     | 359/809 [07:00<08:55,  1.19s/it]
-     69/300      13.9G      1.181     0.4288     0.1287       1185        640:  44%|████▍     | 360/809 [07:02<09:45,  1.30s/it]
-     69/300      13.9G      1.181     0.4288     0.1287        358        640:  45%|████▍     | 362/809 [07:04<09:19,  1.25s/it]
-     69/300      13.9G      1.181     0.4287     0.1287        160        640:  45%|████▍     | 363/809 [07:05<09:09,  1.23s/it]
-     69/300      13.9G       1.18     0.4289     0.1285        280        640:  45%|████▌     | 365/809 [07:08<08:53,  1.20s/it]
-     69/300      13.9G      1.181     0.4285     0.1287        222        640:  45%|████▌     | 367/809 [07:10<08:39,  1.18s/it]
-     69/300      13.9G      1.181     0.4288     0.1287        187        640:  45%|████▌     | 368/809 [07:11<08:37,  1.17s/it]
-     69/300      13.9G      1.185     0.4277      0.131        560        640:  46%|████▌     | 370/809 [07:14<09:09,  1.25s/it]
-     69/300      13.9G      1.189     0.4266     0.1329        588        640:  46%|████▌     | 372/809 [07:16<09:02,  1.24s/it]
-     69/300      13.9G       1.19     0.4264     0.1329        443        640:  46%|████▌     | 373/809 [07:17<08:46,  1.21s/it]
-     69/300      13.9G      1.191     0.4261     0.1328        197        640:  46%|████▋     | 375/809 [07:20<08:29,  1.17s/it]
-     69/300      13.9G       1.19     0.4263     0.1327        198        640:  47%|████▋     | 377/809 [07:22<08:35,  1.19s/it]
-     69/300      13.9G      1.191     0.4261     0.1326        283        640:  47%|████▋     | 378/809 [07:23<08:33,  1.19s/it]
-     69/300      13.9G      1.192     0.4253     0.1327        213        640:  47%|████▋     | 380/809 [07:26<08:38,  1.21s/it]
-     69/300      13.9G      1.193     0.4251     0.1326        180        640:  47%|████▋     | 382/809 [07:28<08:24,  1.18s/it]
-     69/300      13.9G      1.195     0.4245     0.1327        206        640:  47%|████▋     | 384/809 [07:30<08:18,  1.17s/it]
-     69/300      13.9G      1.195     0.4243     0.1327        275        640:  48%|████▊     | 385/809 [07:32<08:27,  1.20s/it]
-     69/300      13.9G      1.194     0.4246     0.1324        236        640:  48%|████▊     | 387/809 [07:34<08:14,  1.17s/it]
-     69/300      13.9G      1.192     0.4253     0.1323        299        640:  48%|████▊     | 389/809 [07:36<08:11,  1.17s/it]
-     69/300      13.9G      1.192     0.4252     0.1323        289        640:  48%|████▊     | 390/809 [07:37<08:05,  1.16s/it]
-     69/300      13.9G      1.191     0.4258      0.132        223        640:  48%|████▊     | 392/809 [07:40<08:00,  1.15s/it]
-     69/300      13.9G      1.192     0.4257     0.1322        341        640:  49%|████▊     | 394/809 [07:42<08:08,  1.18s/it]
-     69/300      13.9G      1.192     0.4258     0.1322        230        640:  49%|████▉     | 396/809 [07:44<08:00,  1.16s/it]
-     69/300      13.9G      1.192      0.426     0.1321        180        640:  49%|████▉     | 397/809 [07:46<08:01,  1.17s/it]
-     69/300      13.9G      1.192     0.4261     0.1321        175        640:  49%|████▉     | 399/809 [07:48<07:57,  1.16s/it]
-     69/300      13.9G       1.19     0.4266     0.1319        164        640:  50%|████▉     | 401/809 [07:50<08:03,  1.19s/it]
-     69/300      13.9G       1.19     0.4269     0.1318        146        640:  50%|████▉     | 402/809 [07:51<07:56,  1.17s/it]
-     69/300      13.9G      1.189      0.427     0.1318        297        640:  50%|████▉     | 404/809 [07:54<08:13,  1.22s/it]
-     69/300      13.9G      1.189     0.4271     0.1316        327        640:  50%|█████     | 406/809 [07:56<07:58,  1.19s/it]
-     69/300      13.9G      1.187     0.4275     0.1315        118        640:  50%|█████     | 407/809 [07:58<08:07,  1.21s/it]
-     69/300      13.9G      1.189     0.4271     0.1318        449        640:  51%|█████     | 409/809 [08:00<08:10,  1.23s/it]
-     69/300      13.9G      1.188     0.4278     0.1318         91        640:  51%|█████     | 411/809 [08:02<07:54,  1.19s/it]
-     69/300      13.9G      1.188     0.4279     0.1317        277        640:  51%|█████     | 412/809 [08:04<07:49,  1.18s/it]
-     69/300      13.9G      1.187     0.4279     0.1315        225        640:  51%|█████     | 414/809 [08:06<07:47,  1.18s/it]
-     69/300      13.9G      1.187     0.4281     0.1313        313        640:  51%|█████▏    | 416/809 [08:08<07:40,  1.17s/it]
-     69/300      13.9G      1.186     0.4282     0.1313        290        640:  52%|█████▏    | 417/809 [08:09<07:41,  1.18s/it]
-     69/300      13.9G      1.188     0.4278     0.1315        308        640:  52%|█████▏    | 419/809 [08:12<07:32,  1.16s/it]
-     69/300      13.9G       1.19     0.4274     0.1318        212        640:  52%|█████▏    | 421/809 [08:14<07:26,  1.15s/it]
-     69/300      13.9G       1.19     0.4273     0.1318        267        640:  52%|█████▏    | 423/809 [08:16<07:35,  1.18s/it]
-     69/300      13.9G      1.191     0.4271     0.1318        445        640:  52%|█████▏    | 424/809 [08:18<07:29,  1.17s/it]
-     69/300      13.9G      1.192     0.4268     0.1318        293        640:  53%|█████▎    | 426/809 [08:20<07:29,  1.17s/it]
-     69/300      13.9G      1.193     0.4265     0.1317        526        640:  53%|█████▎    | 428/809 [08:22<07:32,  1.19s/it]
-     69/300      13.9G      1.192     0.4268     0.1316        160        640:  53%|█████▎    | 429/809 [08:23<07:25,  1.17s/it]
-     69/300      13.9G      1.191     0.4268     0.1315        344        640:  53%|█████▎    | 431/809 [08:26<07:31,  1.19s/it]
-     69/300      13.9G      1.192     0.4268     0.1315        367        640:  54%|█████▎    | 433/809 [08:28<07:32,  1.20s/it]
-     69/300      13.9G      1.192     0.4266     0.1315        326        640:  54%|█████▎    | 434/809 [08:29<07:23,  1.18s/it]
-     69/300      13.9G      1.192     0.4266     0.1315        125        640:  54%|█████▍    | 436/809 [08:32<07:20,  1.18s/it]
-     69/300      13.9G      1.192     0.4264     0.1315        193        640:  54%|█████▍    | 438/809 [08:34<07:12,  1.16s/it]
-     69/300      13.9G      1.194     0.4262     0.1316        242        640:  54%|█████▍    | 440/809 [08:36<07:11,  1.17s/it]
-     69/300      13.9G      1.194     0.4259     0.1316        317        640:  55%|█████▍    | 441/809 [08:38<07:19,  1.19s/it]
-     69/300      13.9G      1.194      0.426     0.1316        485        640:  55%|█████▍    | 443/809 [08:40<07:25,  1.22s/it]
-     69/300      13.9G      1.193     0.4263     0.1317        327        640:  55%|█████▌    | 445/809 [08:42<07:06,  1.17s/it]
-     69/300      13.9G      1.192     0.4267     0.1315        107        640:  55%|█████▌    | 446/809 [08:44<07:01,  1.16s/it]
-     69/300      13.9G      1.192     0.4269     0.1313        207        640:  55%|█████▌    | 448/809 [08:46<07:02,  1.17s/it]
-     69/300      13.9G      1.192     0.4267     0.1312        263        640:  56%|█████▌    | 450/809 [08:48<07:05,  1.18s/it]
-     69/300      13.9G      1.192     0.4265     0.1311        103        640:  56%|█████▌    | 451/809 [08:50<07:14,  1.21s/it]
-     69/300      13.9G      1.192     0.4265      0.131         81        640:  56%|█████▌    | 453/809 [08:52<07:01,  1.18s/it]
-     69/300      13.9G      1.193     0.4261      0.131         94        640:  56%|█████▌    | 455/809 [08:54<06:57,  1.18s/it]
-     69/300      13.9G      1.194     0.4259      0.131        214        640:  56%|█████▋    | 456/809 [08:55<06:51,  1.17s/it]
-     69/300      13.9G      1.193     0.4259     0.1309        472        640:  57%|█████▋    | 458/809 [08:58<06:54,  1.18s/it]
-     69/300      13.9G      1.193      0.426     0.1309        282        640:  57%|█████▋    | 460/809 [09:00<06:43,  1.16s/it]
-     69/300      13.9G      1.192     0.4263     0.1308         95        640:  57%|█████▋    | 462/809 [09:02<06:38,  1.15s/it]
-     69/300      13.9G      1.191     0.4265     0.1307        151        640:  57%|█████▋    | 463/809 [09:04<06:53,  1.19s/it]
-     69/300      13.9G      1.191     0.4265     0.1305        243        640:  57%|█████▋    | 465/809 [09:06<06:47,  1.19s/it]
-     69/300      13.9G      1.191     0.4263     0.1305        233        640:  58%|█████▊    | 467/809 [09:08<06:37,  1.16s/it]
-     69/300      13.9G      1.192      0.426     0.1307        371        640:  58%|█████▊    | 469/809 [09:11<06:34,  1.16s/it]
-     69/300      13.9G      1.192     0.4261     0.1306        199        640:  58%|█████▊    | 470/809 [09:12<06:36,  1.17s/it]
-     69/300      13.9G      1.192      0.426     0.1305        349        640:  58%|█████▊    | 472/809 [09:14<06:29,  1.16s/it]
-     69/300      13.9G       1.19     0.4261     0.1303        180        640:  59%|█████▊    | 474/809 [09:16<06:25,  1.15s/it]
-     69/300      13.9G       1.19     0.4263     0.1302        190        640:  59%|█████▊    | 475/809 [09:18<06:25,  1.15s/it]
-     69/300      13.9G      1.189     0.4265     0.1301        237        640:  59%|█████▉    | 477/809 [09:20<06:28,  1.17s/it]
-     69/300      13.9G      1.188     0.4267     0.1299        196        640:  59%|█████▉    | 479/809 [09:22<06:18,  1.15s/it]
-     69/300      13.9G      1.188      0.427     0.1297        112        640:  59%|█████▉    | 481/809 [09:25<06:25,  1.18s/it]
-     69/300      13.9G      1.187     0.4271     0.1297         97        640:  60%|█████▉    | 482/809 [09:26<06:32,  1.20s/it]
-     69/300      13.9G      1.186     0.4276     0.1295        102        640:  60%|█████▉    | 484/809 [09:28<06:22,  1.18s/it]
-     69/300      13.9G      1.186     0.4277     0.1294        226        640:  60%|██████    | 486/809 [09:31<06:25,  1.19s/it]
-     69/300      13.9G      1.185     0.4278     0.1294         48        640:  60%|██████    | 487/809 [09:32<06:19,  1.18s/it]
-     69/300      13.9G      1.185     0.4279     0.1292        139        640:  60%|██████    | 489/809 [09:34<06:21,  1.19s/it]
-     69/300      13.9G      1.184     0.4283      0.129        374        640:  61%|██████    | 491/809 [09:36<06:10,  1.17s/it]
-     69/300      13.9G      1.182     0.4287     0.1289        348        640:  61%|██████    | 493/809 [09:39<06:04,  1.15s/it]
-     69/300      13.9G      1.183     0.4286      0.129        288        640:  61%|██████    | 494/809 [09:40<06:05,  1.16s/it]
-     69/300      13.9G      1.182     0.4289     0.1289        140        640:  61%|██████▏   | 496/809 [09:42<06:02,  1.16s/it]
-     69/300      13.9G      1.181     0.4292     0.1287         94        640:  62%|██████▏   | 498/809 [09:45<06:02,  1.16s/it]
-     69/300      13.9G      1.181     0.4291     0.1286        235        640:  62%|██████▏   | 499/809 [09:46<06:10,  1.20s/it]
-     69/300      13.9G      1.181     0.4291     0.1288        463        640:  62%|██████▏   | 501/809 [09:48<06:05,  1.19s/it]
-     69/300      13.9G      1.181     0.4293     0.1288        331        640:  62%|██████▏   | 503/809 [09:50<05:57,  1.17s/it]
-     69/300      13.9G      1.182     0.4292     0.1288        360        640:  62%|██████▏   | 504/809 [09:52<05:52,  1.16s/it]
-     69/300      13.9G       1.18     0.4296     0.1286        284        640:  63%|██████▎   | 506/809 [09:54<05:55,  1.17s/it]
-     69/300      13.9G      1.182     0.4292     0.1292        510        640:  63%|██████▎   | 508/809 [09:56<05:51,  1.17s/it]
-     69/300      13.9G      1.181     0.4295     0.1292        206        640:  63%|██████▎   | 510/809 [09:59<05:45,  1.16s/it]
-     69/300      13.9G      1.182     0.4292     0.1298        596        640:  63%|██████▎   | 511/809 [10:00<05:49,  1.17s/it]
-     69/300      13.9G      1.181     0.4293     0.1297        158        640:  63%|██████▎   | 513/809 [10:02<05:52,  1.19s/it]
-     69/300      13.9G      1.181     0.4293     0.1296        400        640:  64%|██████▎   | 515/809 [10:05<05:43,  1.17s/it]
-     69/300      13.9G      1.182     0.4293     0.1297        351        640:  64%|██████▍   | 516/809 [10:06<05:39,  1.16s/it]
-     69/300      13.9G      1.181     0.4296     0.1295        279        640:  64%|██████▍   | 518/809 [10:08<05:36,  1.16s/it]
-     69/300      13.9G       1.18     0.4299     0.1293        223        640:  64%|██████▍   | 520/809 [10:10<05:32,  1.15s/it]
-     69/300      13.9G      1.181     0.4297     0.1295        360        640:  65%|██████▍   | 522/809 [10:13<05:35,  1.17s/it]
-     69/300      13.9G       1.18     0.4299     0.1294        212        640:  65%|██████▍   | 523/809 [10:14<05:32,  1.16s/it]
-     69/300      13.9G      1.179     0.4301     0.1293        258        640:  65%|██████▍   | 525/809 [10:16<05:27,  1.15s/it]
-     69/300      13.9G      1.179     0.4301     0.1292        221        640:  65%|██████▌   | 527/809 [10:18<05:27,  1.16s/it]
-     69/300      13.9G      1.179     0.4302     0.1291        309        640:  65%|██████▌   | 529/809 [10:21<05:25,  1.16s/it]
-     69/300      13.9G      1.179     0.4303      0.129        283        640:  66%|██████▌   | 530/809 [10:22<05:24,  1.16s/it]
-     69/300      13.9G      1.178     0.4305      0.129        344        640:  66%|██████▌   | 532/809 [10:24<05:20,  1.16s/it]
-     69/300      13.9G      1.178     0.4305      0.129        360        640:  66%|██████▌   | 534/809 [10:27<05:15,  1.15s/it]
-     69/300      13.9G      1.178     0.4305      0.129        222        640:  66%|██████▋   | 536/809 [10:29<05:12,  1.14s/it]
-     69/300      13.9G      1.177     0.4306     0.1289        305        640:  66%|██████▋   | 537/809 [10:30<05:14,  1.16s/it]
-     69/300      13.9G      1.177     0.4308     0.1288        215        640:  67%|██████▋   | 539/809 [10:32<05:11,  1.15s/it]
-     69/300      13.9G      1.177     0.4308      0.129        280        640:  67%|██████▋   | 541/809 [10:35<05:16,  1.18s/it]
-     69/300      13.9G      1.176     0.4311     0.1289        200        640:  67%|██████▋   | 542/809 [10:36<05:13,  1.17s/it]
-     69/300      13.9G      1.176     0.4311     0.1289        160        640:  67%|██████▋   | 544/809 [10:38<05:07,  1.16s/it]
-     69/300      13.9G      1.175     0.4312     0.1288        132        640:  67%|██████▋   | 546/809 [10:40<05:05,  1.16s/it]
-     69/300      13.9G      1.175     0.4312     0.1288        331        640:  68%|██████▊   | 548/809 [10:43<05:05,  1.17s/it]
-     69/300      13.9G      1.174     0.4314     0.1287        192        640:  68%|██████▊   | 549/809 [10:44<05:02,  1.16s/it]
-     69/300      13.9G      1.174     0.4316     0.1286        292        640:  68%|██████▊   | 551/809 [10:46<04:58,  1.16s/it]
-     69/300      13.9G      1.173     0.4318     0.1284        368        640:  68%|██████▊   | 553/809 [10:49<04:59,  1.17s/it]
-     69/300      13.9G      1.174     0.4316     0.1288        641        640:  68%|██████▊   | 554/809 [10:50<05:07,  1.21s/it]
-     69/300      13.9G      1.173     0.4318     0.1286        232        640:  69%|██████▊   | 556/809 [10:52<05:11,  1.23s/it]
-     69/300      13.9G      1.173     0.4319     0.1286        451        640:  69%|██████▉   | 558/809 [10:55<05:03,  1.21s/it]
-     69/300      13.9G      1.173     0.4318     0.1287        335        640:  69%|██████▉   | 559/809 [10:56<04:57,  1.19s/it]
-     69/300      13.9G      1.172     0.4321     0.1286        131        640:  69%|██████▉   | 561/809 [10:58<04:55,  1.19s/it]
-     69/300      13.9G      1.171     0.4324     0.1285        158        640:  70%|██████▉   | 563/809 [11:01<04:45,  1.16s/it]
-     69/300      13.9G       1.17     0.4326     0.1284        192        640:  70%|██████▉   | 565/809 [11:03<04:39,  1.15s/it]
-     69/300      13.9G       1.17     0.4326     0.1283        274        640:  70%|██████▉   | 566/809 [11:04<04:38,  1.15s/it]
-     69/300      13.9G       1.17     0.4326     0.1283        253        640:  70%|███████   | 568/809 [11:06<04:38,  1.16s/it]
-     69/300      13.9G       1.17     0.4326     0.1283        288        640:  70%|███████   | 570/809 [11:09<04:40,  1.17s/it]
-     69/300      13.9G       1.17     0.4327     0.1282        138        640:  71%|███████   | 571/809 [11:10<04:42,  1.19s/it]
-     69/300      13.9G      1.169     0.4329     0.1281        173        640:  71%|███████   | 573/809 [11:12<04:37,  1.18s/it]
-     69/300      13.9G      1.169      0.433     0.1281         51        640:  71%|███████   | 575/809 [11:15<04:31,  1.16s/it]
-     69/300      13.9G      1.169     0.4332      0.128        186        640:  71%|███████▏  | 577/809 [11:17<04:32,  1.17s/it]
-     69/300      13.9G       1.17      0.433     0.1281        474        640:  71%|███████▏  | 578/809 [11:18<04:31,  1.17s/it]
-     69/300      13.9G      1.171     0.4326      0.128        264        640:  72%|███████▏  | 580/809 [11:20<04:25,  1.16s/it]
-     69/300      13.9G      1.171     0.4325      0.128        340        640:  72%|███████▏  | 582/809 [11:23<04:19,  1.14s/it]
-     69/300      13.9G      1.171     0.4324      0.128        210        640:  72%|███████▏  | 584/809 [11:25<04:18,  1.15s/it]
-     69/300      13.9G       1.17     0.4326      0.128        117        640:  72%|███████▏  | 585/809 [11:26<04:23,  1.18s/it]
-     69/300      13.9G       1.17     0.4328     0.1278        179        640:  73%|███████▎  | 587/809 [11:29<04:20,  1.17s/it]
-     69/300      13.9G       1.17     0.4329     0.1277        216        640:  73%|███████▎  | 589/809 [11:31<04:13,  1.15s/it]
-     69/300      13.9G       1.17     0.4328     0.1277        362        640:  73%|███████▎  | 590/809 [11:32<04:14,  1.16s/it]
-     69/300      13.9G       1.17     0.4328     0.1277        406        640:  73%|███████▎  | 592/809 [11:34<04:13,  1.17s/it]
-     69/300      13.9G      1.169     0.4332     0.1276        134        640:  73%|███████▎  | 594/809 [11:37<04:12,  1.17s/it]
-     69/300      13.9G      1.169     0.4334     0.1275        192        640:  74%|███████▎  | 595/809 [11:38<04:15,  1.19s/it]
-     69/300      13.9G      1.169     0.4334     0.1281        208        640:  74%|███████▍  | 597/809 [11:40<04:10,  1.18s/it]
-     69/300      13.9G      1.169     0.4333      0.128        276        640:  74%|███████▍  | 599/809 [11:43<04:04,  1.16s/it]
-     69/300      13.9G       1.17     0.4332     0.1281        366        640:  74%|███████▍  | 601/809 [11:45<04:05,  1.18s/it]
-     69/300      13.9G       1.17     0.4333      0.128        147        640:  74%|███████▍  | 602/809 [11:46<04:01,  1.17s/it]
-     69/300      13.9G      1.169     0.4334     0.1279        228        640:  75%|███████▍  | 604/809 [11:48<04:00,  1.18s/it]
-     69/300      13.9G      1.169     0.4336     0.1279        437        640:  75%|███████▍  | 606/809 [11:51<03:57,  1.17s/it]
-     69/300      13.9G      1.169     0.4336      0.128        446        640:  75%|███████▌  | 608/809 [11:53<03:51,  1.15s/it]
-     69/300      13.9G      1.169     0.4338     0.1278        153        640:  75%|███████▌  | 609/809 [11:54<03:51,  1.16s/it]
-     69/300      13.9G      1.169     0.4338     0.1278        280        640:  76%|███████▌  | 611/809 [11:57<03:47,  1.15s/it]
-     69/300      13.9G      1.169     0.4338     0.1277        340        640:  76%|███████▌  | 613/809 [11:59<03:45,  1.15s/it]
-     69/300      13.9G      1.169     0.4338     0.1278        343        640:  76%|███████▌  | 614/809 [12:00<03:52,  1.19s/it]
-     69/300      13.9G       1.17     0.4337      0.128        475        640:  76%|███████▌  | 616/809 [12:02<03:48,  1.18s/it]
-     69/300      13.9G       1.17     0.4335     0.1282        406        640:  76%|███████▋  | 618/809 [12:05<03:44,  1.18s/it]
-     69/300      13.9G       1.17     0.4336      0.128        331        640:  77%|███████▋  | 620/809 [12:07<03:42,  1.18s/it]
-     69/300      13.9G       1.17     0.4336      0.128        267        640:  77%|███████▋  | 621/809 [12:08<03:42,  1.18s/it]
-     69/300      13.9G       1.17     0.4335      0.128        290        640:  77%|███████▋  | 623/809 [12:11<03:37,  1.17s/it]
-     69/300      13.9G      1.169     0.4338     0.1278        146        640:  77%|██��████▋  | 625/809 [12:13<03:35,  1.17s/it]
-     69/300      13.9G      1.168     0.4339     0.1277        236        640:  77%|███████▋  | 626/809 [12:14<03:32,  1.16s/it]
-     69/300      13.9G      1.168      0.434     0.1277        253        640:  78%|███████▊  | 628/809 [12:16<03:27,  1.15s/it]
-     69/300      13.9G      1.168      0.434     0.1276        186        640:  78%|███████▊  | 630/809 [12:19<03:27,  1.16s/it]
-     69/300      13.9G      1.168     0.4339     0.1276        298        640:  78%|███████▊  | 632/809 [12:21<03:22,  1.14s/it]
-     69/300      13.9G      1.167     0.4341     0.1276        222        640:  78%|███████▊  | 633/809 [12:22<03:23,  1.16s/it]
-     69/300      13.9G      1.167     0.4341     0.1277        378        640:  78%|███████▊  | 635/809 [12:24<03:20,  1.15s/it]
-     69/300      13.9G      1.167     0.4341     0.1277        128        640:  79%|███████▊  | 637/809 [12:27<03:17,  1.15s/it]
-     69/300      13.9G      1.167     0.4342     0.1277        232        640:  79%|███████▉  | 639/809 [12:29<03:14,  1.14s/it]
-     69/300      13.9G      1.166     0.4344     0.1276        117        640:  79%|███████▉  | 640/809 [12:30<03:12,  1.14s/it]
-     69/300      13.9G      1.166     0.4346     0.1275        247        640:  79%|███████▉  | 642/809 [12:33<03:16,  1.18s/it]
-     69/300      13.9G      1.167     0.4345     0.1277        215        640:  80%|███████▉  | 644/809 [12:35<03:11,  1.16s/it]
-     69/300      13.9G      1.166     0.4345     0.1276        266        640:  80%|███████▉  | 646/809 [12:37<03:08,  1.16s/it]
-     69/300      13.9G      1.166     0.4345     0.1276        354        640:  80%|███████▉  | 647/809 [12:38<03:07,  1.16s/it]
-     69/300      13.9G      1.167     0.4344     0.1277        281        640:  80%|████████  | 649/809 [12:41<03:14,  1.22s/it]
-     69/300      13.9G      1.166     0.4347     0.1276        465        640:  80%|████████  | 651/809 [12:43<03:08,  1.20s/it]
-     69/300      13.9G      1.166     0.4347     0.1276        244        640:  81%|████████  | 652/809 [12:44<03:05,  1.18s/it]
-     69/300      13.9G      1.165      0.435     0.1275        283        640:  81%|████████  | 654/809 [12:47<03:04,  1.19s/it]
-     69/300      13.9G      1.164     0.4354     0.1274         94        640:  81%|████████  | 656/809 [12:49<03:04,  1.21s/it]
-     69/300      13.9G      1.164     0.4354     0.1273        201        640:  81%|████████  | 657/809 [12:51<03:05,  1.22s/it]
-     69/300      13.9G      1.164     0.4354     0.1275        273        640:  81%|████████▏ | 659/809 [12:53<02:58,  1.19s/it]
-     69/300      13.9G      1.164     0.4354     0.1279        405        640:  82%|████████▏ | 661/809 [12:55<02:55,  1.19s/it]
-     69/300      13.9G      1.164     0.4354     0.1279        326        640:  82%|████████▏ | 662/809 [12:56<02:54,  1.18s/it]
-     69/300      13.9G      1.164     0.4356     0.1278        241        640:  82%|████████▏ | 664/809 [12:59<02:49,  1.17s/it]
-     69/300      13.9G      1.165     0.4353     0.1279        263        640:  82%|████████▏ | 666/809 [13:01<02:48,  1.18s/it]
-     69/300      13.9G      1.165     0.4352      0.128        325        640:  82%|████████▏ | 667/809 [13:02<02:47,  1.18s/it]
-     69/300      13.9G      1.164     0.4353     0.1278        299        640:  83%|████████▎ | 669/809 [13:05<02:43,  1.16s/it]
-     69/300      13.9G      1.165     0.4353      0.128        405        640:  83%|████████▎ | 671/809 [13:07<02:38,  1.15s/it]
-     69/300      13.9G      1.164     0.4355      0.128        175        640:  83%|████████▎ | 673/809 [13:09<02:37,  1.16s/it]
-     69/300      13.9G      1.164     0.4357     0.1279        155        640:  83%|████████▎ | 674/809 [13:10<02:34,  1.15s/it]
-     69/300      13.9G      1.164     0.4357     0.1279        310        640:  84%|████████▎ | 676/809 [13:13<02:38,  1.19s/it]
-     69/300      13.9G      1.163     0.4359     0.1278        239        640:  84%|████████▍ | 678/809 [13:15<02:33,  1.17s/it]
-     69/300      13.9G      1.162      0.436     0.1277        141        640:  84%|████████▍ | 679/809 [13:16<02:37,  1.21s/it]
-     69/300      13.9G      1.162      0.436     0.1277        298        640:  84%|████████▍ | 681/809 [13:19<02:32,  1.19s/it]
-     69/300      13.9G      1.162     0.4361     0.1277        434        640:  84%|████████▍ | 683/809 [13:21<02:26,  1.16s/it]
-     69/300      13.9G      1.161     0.4364     0.1277        112        640:  85%|████████▍ | 685/809 [13:23<02:24,  1.16s/it]
-     69/300      13.9G      1.161     0.4364     0.1276        325        640:  85%|████████▍ | 686/809 [13:24<02:22,  1.16s/it]
-     69/300      13.9G      1.161     0.4364     0.1276        302        640:  85%|████████▌ | 688/809 [13:27<02:19,  1.15s/it]
-     69/300      13.9G      1.161     0.4365     0.1275        310        640:  85%|████████▌ | 690/809 [13:29<02:18,  1.16s/it]
-     69/300      13.9G       1.16     0.4369     0.1273        145        640:  86%|████████▌ | 692/809 [13:31<02:15,  1.16s/it]
-     69/300      13.9G       1.16      0.437     0.1273         86        640:  86%|████████▌ | 693/809 [13:33<02:14,  1.16s/it]
-     69/300      13.9G      1.159     0.4372     0.1272        167        640:  86%|████████▌ | 695/809 [13:35<02:19,  1.23s/it]
-     69/300      13.9G      1.159     0.4373     0.1271        207        640:  86%|████████▌ | 696/809 [13:36<02:15,  1.20s/it]
-     69/300      13.9G      1.158     0.4376      0.127        127        640:  86%|████████▋ | 698/809 [13:39<02:15,  1.22s/it]
-     69/300      13.9G      1.158     0.4376     0.1268        247        640:  87%|████████▋ | 700/809 [13:41<02:09,  1.19s/it]
-     69/300      13.9G      1.158     0.4377     0.1267        223        640:  87%|████████▋ | 702/809 [13:43<02:07,  1.19s/it]
-     69/300      13.9G      1.158     0.4375     0.1268        406        640:  87%|████████▋ | 703/809 [13:45<02:06,  1.19s/it]
-     69/300      13.9G      1.159     0.4373     0.1272        375        640:  87%|████████▋ | 705/809 [13:47<02:04,  1.20s/it]
-     69/300      13.9G       1.16      0.437     0.1274        249        640:  87%|████████▋ | 707/809 [13:49<02:00,  1.18s/it]
-     69/300      13.9G      1.161     0.4368     0.1279        590        640:  88%|████████▊ | 708/809 [13:50<01:59,  1.18s/it]
-     69/300      13.9G      1.161     0.4368     0.1278        492        640:  88%|████████▊ | 710/809 [13:53<01:55,  1.16s/it]
-     69/300      13.9G      1.161     0.4367     0.1278        438        640:  88%|████████▊ | 712/809 [13:55<01:55,  1.19s/it]
-     69/300      13.9G      1.161     0.4367     0.1277        234        640:  88%|████████▊ | 713/809 [13:56<01:55,  1.20s/it]
-     69/300      13.9G       1.16     0.4369     0.1277        327        640:  88%|████████▊ | 715/809 [13:59<01:50,  1.17s/it]
-     69/300      13.9G       1.16     0.4369     0.1276        302        640:  89%|████████▊ | 717/809 [14:01<01:48,  1.17s/it]
-     69/300      13.9G      1.161     0.4367     0.1277        392        640:  89%|████████▉ | 719/809 [14:03<01:45,  1.17s/it]
-     69/300      13.9G      1.161     0.4367     0.1277        354        640:  89%|████████▉ | 720/809 [14:04<01:42,  1.16s/it]
-     69/300      13.9G      1.161     0.4366     0.1276        338        640:  89%|████████▉ | 722/809 [14:07<01:46,  1.22s/it]
-     69/300      13.9G       1.16     0.4369     0.1275        106        640:  89%|████████▉ | 724/809 [14:09<01:42,  1.21s/it]
-     69/300      13.9G      1.159      0.437     0.1274         70        640:  90%|████████▉ | 725/809 [14:11<01:41,  1.21s/it]
-     69/300      13.9G      1.158     0.4373     0.1273        202        640:  90%|████████▉ | 727/809 [14:13<01:38,  1.20s/it]
-     69/300      13.9G      1.158     0.4374     0.1272        229        640:  90%|█████████ | 729/809 [14:15<01:37,  1.21s/it]
-     69/300      13.9G      1.158     0.4375     0.1272        251        640:  90%|█████████ | 730/809 [14:17<01:35,  1.21s/it]
-     69/300      13.9G      1.157     0.4376     0.1272        157        640:  90%|█████████ | 732/809 [14:19<01:34,  1.23s/it]
-     69/300      13.9G      1.157     0.4378     0.1271        248        640:  91%|█████████ | 734/809 [14:22<01:30,  1.20s/it]
-     69/300      13.9G      1.157     0.4377     0.1271        307        640:  91%|█████████ | 735/809 [14:23<01:27,  1.19s/it]
-     69/300      13.9G      1.157     0.4377     0.1271        207        640:  91%|█████████ | 737/809 [14:25<01:25,  1.19s/it]
-     69/300      13.9G      1.157     0.4378     0.1271        309        640:  91%|█████████▏| 739/809 [14:27<01:21,  1.17s/it]
-     69/300      13.9G      1.156      0.438      0.127        171        640:  91%|█████████▏| 740/809 [14:28<01:20,  1.17s/it]
-     69/300      13.9G      1.156      0.438      0.127        298        640:  92%|█████████▏| 742/809 [14:31<01:17,  1.16s/it]
-     69/300      13.9G      1.155     0.4381     0.1269        155        640:  92%|█████████▏| 744/809 [14:33<01:15,  1.16s/it]
-     69/300      13.9G      1.154     0.4384     0.1268        233        640:  92%|█████████▏| 746/809 [14:35<01:13,  1.17s/it]
-     69/300      13.9G      1.154     0.4386     0.1268        220        640:  92%|█████████▏| 747/809 [14:37<01:12,  1.17s/it]
-     69/300      13.9G      1.153     0.4386     0.1267        309        640:  93%|█████████▎| 749/809 [14:39<01:09,  1.16s/it]
-     69/300      13.9G      1.153     0.4388     0.1266        247        640:  93%|█████████▎| 751/809 [14:41<01:07,  1.16s/it]
-     69/300      13.9G      1.152      0.439     0.1266        190        640:  93%|█████████▎| 752/809 [14:42<01:06,  1.16s/it]
-     69/300      13.9G      1.152     0.4391     0.1265        389        640:  93%|█████████▎| 754/809 [14:45<01:04,  1.17s/it]
-     69/300      13.9G      1.152     0.4391     0.1265        284        640:  93%|█████████▎| 756/809 [14:47<01:01,  1.17s/it]
-     69/300      13.9G      1.153     0.4391     0.1264        200        640:  94%|█████████▎| 758/809 [14:50<01:00,  1.19s/it]
-     69/300      13.9G      1.153      0.439     0.1264        313        640:  94%|█████████▍| 759/809 [14:51<00:59,  1.19s/it]
-     69/300      13.9G      1.154     0.4387     0.1265        264        640:  94%|█████████▍| 761/809 [14:53<00:56,  1.18s/it]
-     69/300      13.9G      1.153     0.4388     0.1265        170        640:  94%|█████████▍| 762/809 [14:54<00:54,  1.17s/it]
-     69/300      13.9G      1.153     0.4388     0.1264         95        640:  94%|█████████▍| 764/809 [14:57<00:52,  1.17s/it]
-     69/300      13.9G      1.152     0.4389     0.1263        227        640:  95%|█████████▍| 766/809 [14:59<00:50,  1.18s/it]
-     69/300      13.9G      1.152     0.4389     0.1262        383        640:  95%|█████████▍| 767/809 [15:00<00:49,  1.17s/it]
-     69/300      13.9G      1.152     0.4389     0.1261        410        640:  95%|█████████▌| 769/809 [15:02<00:47,  1.18s/it]
-     69/300      13.9G      1.153     0.4388     0.1261        351        640:  95%|█████████▌| 771/809 [15:05<00:43,  1.15s/it]
-     69/300      13.9G      1.152     0.4389     0.1261        251        640:  96%|█████████▌| 773/809 [15:07<00:41,  1.15s/it]
-     69/300      13.9G      1.152     0.4391      0.126        116        640:  96%|█████████▌| 774/809 [15:08<00:40,  1.15s/it]
-     69/300      13.9G      1.152     0.4393     0.1259        121        640:  96%|█████████▌| 776/809 [15:10<00:37,  1.15s/it]
-     69/300      13.9G      1.152     0.4394     0.1259        212        640:  96%|█████████▌| 778/809 [15:13<00:36,  1.16s/it]
-     69/300      13.9G      1.153     0.4392      0.126        240        640:  96%|█████████▋| 780/809 [15:15<00:33,  1.16s/it]
-     69/300      13.9G      1.152     0.4394      0.126        151        640:  97%|█████████▋| 781/809 [15:16<00:32,  1.17s/it]
-     69/300      13.9G      1.152     0.4394     0.1259        109        640:  97%|█████████▋| 783/809 [15:19<00:30,  1.17s/it]
-     69/300      13.9G      1.152     0.4393     0.1259        250        640:  97%|█████████▋| 784/809 [15:20<00:29,  1.19s/it]
-     69/300      13.9G      1.152     0.4393     0.1259        117        640:  97%|█████████▋| 786/809 [15:22<00:27,  1.22s/it]
-     69/300      13.9G      1.152     0.4395     0.1258        133        640:  97%|█████████▋| 788/809 [15:25<00:25,  1.20s/it]
-     69/300      13.9G      1.151     0.4396     0.1257        244        640:  98%|█████████▊| 790/809 [15:27<00:22,  1.20s/it]
-     69/300      13.9G      1.151     0.4396     0.1257        293        640:  98%|█████████▊| 791/809 [15:28<00:21,  1.19s/it]
-     69/300      13.9G       1.15     0.4399     0.1256        195        640:  98%|█████████▊| 793/809 [15:31<00:18,  1.18s/it]
-     69/300      13.9G       1.15     0.4399     0.1255        234        640:  98%|█████████▊| 795/809 [15:33<00:16,  1.15s/it]
-     69/300      13.9G       1.15       0.44     0.1255        146        640:  98%|█████████▊| 796/809 [15:34<00:14,  1.15s/it]
-     69/300      13.9G       1.15       0.44     0.1254        205        640:  99%|█████████▊| 798/809 [15:36<00:12,  1.15s/it]
-     69/300      13.9G       1.15     0.4399     0.1255        374        640:  99%|█████████▉| 800/809 [15:39<00:10,  1.14s/it]
-     69/300      13.9G       1.15     0.4401     0.1254        158        640:  99%|█████████▉| 802/809 [15:41<00:08,  1.15s/it]
-     69/300      13.9G       1.15     0.4401     0.1254        249        640:  99%|█████████▉| 803/809 [15:42<00:06,  1.15s/it]
-     69/300      13.9G       1.15       0.44     0.1254        238        640: 100%|█████████▉| 805/809 [15:44<00:04,  1.15s/it]
-     69/300      13.9G       1.15     0.4401     0.1252        130        640: 100%|█████████▉| 807/809 [15:47<00:02,  1.16s/it]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   0%|          | 0/69 [00:00<?, ?it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  12%|█▏        | 8/69 [00:02<00:15,  3.96it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  23%|██▎       | 16/69 [00:04<00:13,  3.98it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  35%|███▍      | 24/69 [00:05<00:11,  4.09it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  46%|████▋     | 32/69 [00:07<00:09,  3.93it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  58%|█████▊    | 40/69 [00:09<00:07,  4.07it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  70%|██████▉   | 48/69 [00:12<00:05,  3.64it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  81%|████████  | 56/69 [00:14<00:03,  4.05it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  93%|█████████▎| 64/69 [00:15<00:01,  4.04it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.04it/s]
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.04it/s]
-                   all        548      38759      0.503      0.356      0.358      0.197
-     70/300      13.2G      1.333     0.3869     0.1169         99        640:   0%|          | 1/809 [00:01<15:53,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.566     0.2988     0.1312        257        640:   0%|          | 3/809 [00:03<15:49,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.57     0.3045      0.157        186        640:   1%|          | 5/809 [00:05<15:43,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.54     0.3112      0.153         61        640:   1%|          | 6/809 [00:07<15:40,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.489     0.3291     0.1488         77        640:   1%|          | 8/809 [00:09<15:53,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.44     0.3429     0.1411        324        640:   1%|          | 10/809 [00:11<15:34,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.39     0.3587     0.1383        171        640:   1%|▏         | 11/809 [00:12<15:29,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.34     0.3705     0.1323        265        640:   2%|▏         | 13/809 [00:15<15:26,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.324     0.3787     0.1297         83        640:   2%|▏         | 15/809 [00:17<15:29,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.303     0.3848     0.1283        260        640:   2%|▏         | 17/809 [00:19<15:23,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.293     0.3879     0.1275        312        640:   2%|▏         | 18/809 [00:21<15:20,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.275     0.3942     0.1268        194        640:   2%|▏         | 20/809 [00:23<15:30,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.268     0.3931     0.1271        341        640:   3%|▎         | 22/809 [00:25<15:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.277     0.3907     0.1304        714        640:   3%|▎         | 23/809 [00:27<15:37,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.254     0.3963     0.1302        306        640:   3%|▎         | 25/809 [00:29<15:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.254      0.398     0.1324        411        640:   3%|▎         | 27/809 [00:31<15:20,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.265     0.3949     0.1367        344        640:   3%|▎         | 28/809 [00:32<15:31,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.244     0.3997     0.1343        206        640:   4%|▎         | 30/809 [00:35<15:17,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.232     0.4044     0.1338        192        640:   4%|▍         | 32/809 [00:37<15:24,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.228     0.4097     0.1332        414        640:   4%|▍         | 34/809 [00:39<15:07,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.229     0.4089     0.1335        231        640:   4%|▍         | 36/809 [00:42<15:09,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.23     0.4104     0.1341        145        640:   5%|▍         | 37/809 [00:43<15:04,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.218      0.415     0.1327        232        640:   5%|▍         | 39/809 [00:45<15:21,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.223     0.4136     0.1341        252        640:   5%|▌         | 41/809 [00:48<15:13,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.214     0.4167     0.1337         83        640:   5%|▌         | 42/809 [00:49<15:19,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.215     0.4161     0.1326        367        640:   5%|▌         | 44/809 [00:51<14:54,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.218     0.4151     0.1326        278        640:   6%|▌         | 46/809 [00:54<14:41,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.223     0.4132     0.1324        247        640:   6%|▌         | 48/809 [00:56<14:51,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.219     0.4139      0.132        222        640:   6%|▌         | 49/809 [00:57<14:45,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.22     0.4126     0.1322        315        640:   6%|▋         | 51/809 [00:59<14:31,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.219     0.4138     0.1319        337        640:   7%|▋         | 53/809 [01:02<14:32,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.223     0.4125     0.1331        234        640:   7%|▋         | 55/809 [01:04<14:38,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.229     0.4111     0.1333        195        640:   7%|▋         | 56/809 [01:05<14:42,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.224     0.4127     0.1335        111        640:   7%|▋         | 58/809 [01:08<14:25,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.226     0.4124     0.1343        295        640:   7%|▋         | 60/809 [01:10<14:16,  1.14s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.224      0.413     0.1341        296        640:   8%|▊         | 62/809 [01:12<14:18,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.216     0.4155     0.1336         93        640:   8%|▊         | 63/809 [01:13<14:13,  1.14s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.22     0.4152     0.1343        240        640:   8%|▊         | 65/809 [01:16<14:14,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.219     0.4149     0.1337        196        640:   8%|▊         | 67/809 [01:18<14:35,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.222     0.4149     0.1348        220        640:   8%|▊         | 68/809 [01:19<14:23,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.223     0.4146     0.1344        325        640:   9%|▊         | 70/809 [01:22<14:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.224     0.4144     0.1343        448        640:   9%|▉         | 72/809 [01:24<14:40,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.224     0.4148     0.1339        313        640:   9%|▉         | 73/809 [01:25<14:21,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.23     0.4126     0.1343        103        640:   9%|▉         | 75/809 [01:27<14:10,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.228      0.413     0.1331        210        640:  10%|▉         | 77/809 [01:30<14:12,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.23     0.4125      0.133        362        640:  10%|▉         | 79/809 [01:32<14:04,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.227     0.4131     0.1325        189        640:  10%|▉         | 80/809 [01:33<14:11,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.224     0.4142     0.1321        212        640:  10%|█         | 82/809 [01:35<13:59,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.218     0.4152     0.1314        156        640:  10%|█         | 84/809 [01:38<13:53,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.22     0.4149      0.132        168        640:  11%|█         | 86/809 [01:40<14:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.219     0.4153     0.1322        194        640:  11%|█         | 87/809 [01:41<13:53,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.219     0.4152     0.1322        485        640:  11%|█         | 89/809 [01:44<14:14,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.212     0.4176     0.1311        227        640:  11%|█         | 91/809 [01:46<13:53,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.211      0.418     0.1307        281        640:  11%|█▏        | 92/809 [01:47<14:04,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.209     0.4184     0.1305        171        640:  12%|█▏        | 94/809 [01:50<14:14,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.204     0.4215     0.1299        110        640:  12%|█▏        | 96/809 [01:52<14:15,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.202     0.4225     0.1296        189        640:  12%|█▏        | 97/809 [01:53<13:57,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G        1.2     0.4227      0.129        140        640:  12%|█▏        | 99/809 [01:55<13:45,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G        1.2     0.4225     0.1289        317        640:  12%|█▏        | 101/809 [01:58<13:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.198     0.4229     0.1288        128        640:  13%|█▎        | 103/809 [02:00<13:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.197     0.4231     0.1288        191        640:  13%|█▎        | 104/809 [02:01<14:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.202     0.4216     0.1291        294        640:  13%|█▎        | 106/809 [02:04<13:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.199     0.4223     0.1285        456        640:  13%|█▎        | 108/809 [02:06<13:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.197     0.4227     0.1283        172        640:  13%|█▎        | 109/809 [02:08<13:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.201      0.422     0.1289        338        640:  14%|█▎        | 111/809 [02:09<13:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.205     0.4199     0.1292        290        640:  14%|█▍        | 113/809 [02:12<13:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.208     0.4195     0.1295        226        640:  14%|█▍        | 115/809 [02:14<13:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.208     0.4193     0.1297        366        640:  14%|█▍        | 116/809 [02:15<13:18,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.207     0.4199      0.129        215        640:  15%|█▍        | 118/809 [02:18<13:09,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.205     0.4207     0.1283        222        640:  15%|█▍        | 120/809 [02:20<13:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.205     0.4209     0.1282        113        640:  15%|█▍        | 121/809 [02:21<13:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.204      0.421     0.1277        269        640:  15%|█▌        | 123/809 [02:24<13:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.205     0.4208     0.1272        197        640:  15%|█▌        | 125/809 [02:26<13:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.201     0.4222     0.1266        174        640:  16%|█▌        | 127/809 [02:28<13:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.203     0.4214     0.1269        459        640:  16%|█▌        | 128/809 [02:30<13:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.203     0.4213     0.1268        414        640:  16%|█▌        | 130/809 [02:32<13:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.203     0.4219     0.1264        139        640:  16%|█▋        | 132/809 [02:34<13:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G        1.2     0.4225      0.126        213        640:  16%|█▋        | 133/809 [02:35<13:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.196     0.4236     0.1254        217        640:  17%|█▋        | 135/809 [02:38<14:32,  1.29s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.195     0.4238     0.1252        116        640:  17%|█▋        | 136/809 [02:39<14:17,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.193     0.4242     0.1249        211        640:  17%|█▋        | 138/809 [02:42<13:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.192     0.4244     0.1245        170        640:  17%|█▋        | 140/809 [02:44<13:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.188     0.4259      0.124        162        640:  18%|█▊        | 142/809 [02:46<12:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.189     0.4257     0.1241        300        640:  18%|█▊        | 143/809 [02:47<12:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.192     0.4251     0.1239        179        640:  18%|█▊        | 145/809 [02:50<13:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.194     0.4245     0.1253        280        640:  18%|█▊        | 147/809 [02:52<13:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.193     0.4245     0.1254        172        640:  18%|█▊        | 148/809 [02:53<13:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.192     0.4254     0.1252        390        640:  19%|█▊        | 150/809 [02:56<12:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.189     0.4266     0.1246        338        640:  19%|█▉        | 152/809 [02:58<12:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.189     0.4269     0.1245        182        640:  19%|█▉        | 153/809 [02:59<12:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.189     0.4267     0.1241        138        640:  19%|█▉        | 155/809 [03:02<12:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.183     0.4279     0.1236        140        640:  19%|█▉        | 157/809 [03:04<12:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4288     0.1235         83        640:  20%|█▉        | 159/809 [03:06<12:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4289     0.1235        396        640:  20%|█▉        | 160/809 [03:07<12:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4293     0.1235        232        640:  20%|██        | 162/809 [03:10<12:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4289     0.1234        299        640:  20%|██        | 164/809 [03:12<12:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4288     0.1234        306        640:  20%|██        | 165/809 [03:13<12:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182      0.429     0.1234        187        640:  21%|██        | 167/809 [03:16<12:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4289     0.1233        264        640:  21%|██        | 169/809 [03:18<12:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4292     0.1231        379        640:  21%|██        | 170/809 [03:19<12:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.179     0.4298     0.1228        232        640:  21%|██▏       | 172/809 [03:22<12:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4303      0.123        108        640:  22%|██▏       | 174/809 [03:24<12:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4313     0.1225        133        640:  22%|██▏       | 176/809 [03:26<12:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4313     0.1225        371        640:  22%|██▏       | 177/809 [03:28<12:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4309     0.1226        227        640:  22%|██▏       | 179/809 [03:30<12:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173     0.4315     0.1222        231        640:  22%|██▏       | 181/809 [03:32<12:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173     0.4316      0.122        411        640:  22%|██▏       | 182/809 [03:33<12:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.169     0.4331     0.1215         76        640:  23%|██▎       | 184/809 [03:36<12:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.167     0.4334     0.1213        269        640:  23%|██▎       | 186/809 [03:38<12:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.166     0.4335     0.1212        301        640:  23%|██▎       | 188/809 [03:40<11:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.163     0.4343     0.1209        102        640:  23%|██▎       | 189/809 [03:42<11:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.164     0.4337      0.121        250        640:  24%|██▎       | 191/809 [03:44<11:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.16     0.4346     0.1208         90        640:  24%|██▍       | 193/809 [03:46<11:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159     0.4349     0.1206        176        640:  24%|██▍       | 195/809 [03:48<11:42,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159      0.435     0.1206        341        640:  24%|██▍       | 196/809 [03:50<11:47,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.156     0.4361     0.1202        333        640:  24%|██▍       | 198/809 [03:52<11:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.155     0.4363     0.1202        221        640:  25%|██▍       | 200/809 [03:54<11:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.156     0.4362     0.1208        280        640:  25%|██▍       | 202/809 [03:57<11:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.155     0.4364     0.1206        114        640:  25%|██▌       | 204/809 [03:59<11:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.155     0.4363     0.1205        315        640:  25%|██▌       | 205/809 [04:00<11:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.152     0.4368     0.1204        118        640:  26%|██▌       | 207/809 [04:03<11:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.152     0.4369     0.1205        409        640:  26%|██▌       | 209/809 [04:05<11:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.15     0.4376     0.1203        132        640:  26%|██▌       | 210/809 [04:06<11:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.149     0.4382     0.1208        230        640:  26%|██▌       | 212/809 [04:08<11:27,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.147     0.4383     0.1207        274        640:  26%|██▋       | 214/809 [04:11<11:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.145     0.4394     0.1203        187        640:  27%|██▋       | 216/809 [04:13<11:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.147     0.4388     0.1202        208        640:  27%|██▋       | 217/809 [04:14<11:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.147     0.4391     0.1201        286        640:  27%|██▋       | 219/809 [04:17<11:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.146     0.4394     0.1199        270        640:  27%|██▋       | 221/809 [04:19<11:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.147     0.4393     0.1199        214        640:  27%|██▋       | 222/809 [04:20<11:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.148     0.4392     0.1199        208        640:  28%|���█▊       | 224/809 [04:23<11:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.146     0.4399     0.1198        113        640:  28%|██▊       | 226/809 [04:25<11:11,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.145     0.4402     0.1198        242        640:  28%|██▊       | 228/809 [04:27<11:08,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.146       0.44     0.1199        309        640:  28%|██▊       | 229/809 [04:28<11:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.145     0.4407     0.1197        344        640:  29%|██▊       | 231/809 [04:31<11:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.147     0.4402     0.1201        263        640:  29%|██▉       | 233/809 [04:33<11:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.145     0.4408     0.1199         82        640:  29%|██▉       | 234/809 [04:34<11:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.142     0.4415     0.1195         85        640:  29%|██▉       | 236/809 [04:37<11:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.141     0.4417     0.1194        134        640:  29%|██▉       | 238/809 [04:39<10:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.141     0.4414     0.1192        252        640:  30%|██▉       | 239/809 [04:40<11:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.141     0.4414     0.1191        224        640:  30%|██▉       | 241/809 [04:42<11:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.142     0.4413     0.1192        216        640:  30%|███       | 243/809 [04:45<10:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.146     0.4402     0.1198        364        640:  30%|███       | 245/809 [04:47<10:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.146     0.4402     0.1198        197        640:  30%|███       | 246/809 [04:48<10:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.149     0.4395     0.1199        165        640:  31%|███       | 248/809 [04:51<11:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.147     0.4401     0.1198        296        640:  31%|███       | 250/809 [04:53<10:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.147     0.4402     0.1198        334        640:  31%|███       | 251/809 [04:54<11:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.146     0.4404     0.1198        384        640:  31%|███▏      | 253/809 [04:56<10:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.144     0.4408     0.1194        172        640:  32%|███▏      | 255/809 [04:59<10:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.143     0.4408     0.1194        205        640:  32%|███▏      | 257/809 [05:01<10:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.141     0.4414     0.1192        118        640:  32%|███▏      | 258/809 [05:02<10:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.141     0.4416     0.1193        315        640:  32%|███▏      | 260/809 [05:05<10:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.139     0.4418     0.1191        137        640:  32%|███▏      | 262/809 [05:07<10:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.139      0.442      0.119        171        640:  33%|███▎      | 263/809 [05:08<10:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.14     0.4417     0.1191        411        640:  33%|███▎      | 265/809 [05:11<10:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.143     0.4408     0.1195        322        640:  33%|███▎      | 267/809 [05:13<10:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.144     0.4403     0.1196         91        640:  33%|███▎      | 269/809 [05:15<10:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.145     0.4401     0.1197        118        640:  33%|███▎      | 270/809 [05:16<10:20,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.148     0.4392     0.1201        182        640:  34%|███▎      | 272/809 [05:19<11:02,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.147     0.4398     0.1198        193        640:  34%|███▍      | 274/809 [05:21<10:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.146     0.4404     0.1197        131        640:  34%|███▍      | 275/809 [05:22<10:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.143     0.4409     0.1196         68        640:  34%|███▍      | 277/809 [05:25<10:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.144     0.4407     0.1202        392        640:  34%|███▍      | 279/809 [05:27<10:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.143     0.4415       0.12        210        640:  35%|███▍      | 280/809 [05:28<10:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.144      0.441     0.1201        214        640:  35%|███▍      | 282/809 [05:31<10:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.144     0.4411     0.1201        141        640:  35%|███▌      | 284/809 [05:33<10:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.147       0.44     0.1206        227        640:  35%|███▌      | 286/809 [05:35<10:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.147     0.4401     0.1207        143        640:  35%|███▌      | 287/809 [05:37<10:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.147     0.4403     0.1211        246        640:  36%|███▌      | 289/809 [05:39<10:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.146     0.4411     0.1209        262        640:  36%|███▌      | 291/809 [05:41<09:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.145     0.4412     0.1209         92        640:  36%|███▌      | 292/809 [05:42<10:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.146     0.4408     0.1211        331        640:  36%|███▋      | 294/809 [05:45<09:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.148     0.4404     0.1214        321        640:  37%|███▋      | 296/809 [05:47<10:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.149     0.4399     0.1218        192        640:  37%|███▋      | 297/809 [05:48<10:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.149     0.4399     0.1217        193        640:  37%|███▋      | 299/809 [05:51<10:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.149     0.4403     0.1217        336        640:  37%|███▋      | 301/809 [05:53<10:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.149     0.4403     0.1218        318        640:  37%|███▋      | 302/809 [05:54<10:20,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.149     0.4402     0.1218        204        640:  38%|███▊      | 304/809 [05:57<10:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.151     0.4394     0.1222        386        640:  38%|███▊      | 306/809 [05:59<10:13,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.151     0.4394     0.1221        274        640:  38%|███▊      | 307/809 [06:01<10:17,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.154     0.4386     0.1227        444        640:  38%|███▊      | 309/809 [06:03<09:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.155     0.4381     0.1227        422        640:  38%|███▊      | 311/809 [06:05<09:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.155     0.4382     0.1226        162        640:  39%|███▊      | 312/809 [06:07<10:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.157     0.4377     0.1233        325        640:  39%|███▉      | 314/809 [06:09<10:12,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.158     0.4375     0.1234        333        640:  39%|███▉      | 316/809 [06:11<09:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.158     0.4375     0.1233        178        640:  39%|███▉      | 317/809 [06:12<09:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.158     0.4375     0.1234        226        640:  39%|███▉      | 319/809 [06:15<09:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.158     0.4373     0.1235        270        640:  40%|███▉      | 321/809 [06:17<09:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159     0.4372     0.1235        244        640:  40%|███▉      | 323/809 [06:19<09:20,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159      0.437     0.1235        343        640:  40%|████      | 324/809 [06:21<09:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159     0.4371     0.1234        123        640:  40%|████      | 326/809 [06:23<09:10,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.16     0.4364     0.1236        424        640:  41%|████      | 328/809 [06:25<09:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159     0.4366     0.1235        107        640:  41%|████      | 329/809 [06:26<09:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159     0.4367     0.1234        229        640:  41%|████      | 331/809 [06:29<09:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.158     0.4368     0.1232        270        640:  41%|████      | 333/809 [06:31<09:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159     0.4367     0.1233        278        640:  41%|████▏     | 335/809 [06:33<09:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159     0.4367     0.1234        232        640:  42%|████▏     | 336/809 [06:35<09:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.157     0.4373     0.1232        158        640:  42%|████▏     | 338/809 [06:37<09:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.158     0.4372     0.1233        318        640:  42%|████▏     | 340/809 [06:39<09:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.158     0.4373     0.1233        298        640:  42%|████▏     | 341/809 [06:41<09:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.156     0.4377     0.1232        168        640:  42%|████▏     | 343/809 [06:43<09:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.157     0.4376     0.1232        265        640:  43%|████▎     | 345/809 [06:45<09:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.157     0.4374     0.1231        290        640:  43%|█��██▎     | 346/809 [06:46<09:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.158      0.437     0.1235        255        640:  43%|████▎     | 348/809 [06:49<09:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159     0.4369     0.1236        411        640:  43%|████▎     | 350/809 [06:51<08:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159     0.4369     0.1236        240        640:  44%|████▎     | 352/809 [06:53<08:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.16     0.4364     0.1236        350        640:  44%|████▍     | 354/809 [06:56<08:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.16     0.4365     0.1236        368        640:  44%|████▍     | 356/809 [06:58<08:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.159     0.4366     0.1235        230        640:  44%|████▍     | 357/809 [06:59<08:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.162     0.4359     0.1243        253        640:  44%|████▍     | 359/809 [07:02<08:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.161     0.4362     0.1243        121        640:  45%|████▍     | 361/809 [07:04<08:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.161     0.4361     0.1242        227        640:  45%|████▍     | 362/809 [07:05<08:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.161     0.4363     0.1241        228        640:  45%|████▍     | 364/809 [07:07<08:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.161     0.4361     0.1243        260        640:  45%|████▌     | 366/809 [07:10<08:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.162     0.4357     0.1244        337        640:  45%|████▌     | 367/809 [07:11<08:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.162     0.4357     0.1243        148        640:  46%|████▌     | 369/809 [07:13<08:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.165     0.4349     0.1253        584        640:  46%|████▌     | 371/809 [07:16<08:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.168      0.434     0.1256        360        640:  46%|████▌     | 373/809 [07:18<08:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.169     0.4338     0.1257        241        640:  46%|████▌     | 374/809 [07:19<08:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.168     0.4339     0.1257        202        640:  46%|████▋     | 376/809 [07:22<08:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.168      0.434     0.1257        305        640:  47%|████▋     | 378/809 [07:24<08:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.168      0.434     0.1257        273        640:  47%|████▋     | 379/809 [07:25<08:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.169     0.4337     0.1256        172        640:  47%|████▋     | 381/809 [07:27<08:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.171     0.4334      0.126        405        640:  47%|████▋     | 383/809 [07:30<08:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.172     0.4329      0.126        174        640:  48%|████▊     | 385/809 [07:32<08:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.171     0.4331     0.1261        187        640:  48%|████▊     | 386/809 [07:33<08:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173     0.4327     0.1268         87        640:  48%|████▊     | 388/809 [07:36<08:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173     0.4324     0.1267        132        640:  48%|████▊     | 390/809 [07:38<08:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4321     0.1267        285        640:  48%|████▊     | 391/809 [07:39<08:00,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4313     0.1273        170        640:  49%|████▊     | 393/809 [07:41<08:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4305     0.1277        312        640:  49%|████▉     | 395/809 [07:44<07:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.185     0.4295     0.1292        474        640:  49%|████▉     | 397/809 [07:46<08:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.186     0.4293     0.1293        201        640:  49%|████▉     | 398/809 [07:47<08:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.187     0.4291     0.1296        124        640:  49%|████▉     | 400/809 [07:50<08:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.186     0.4294     0.1295        279        640:  50%|████▉     | 402/809 [07:52<07:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.186     0.4294     0.1295        472        640:  50%|████▉     | 403/809 [07:53<07:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.187     0.4295     0.1295        207        640:  50%|█████     | 405/809 [07:55<07:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.187     0.4294     0.1298        374        640:  50%|█████     | 407/809 [07:58<07:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.188     0.4295     0.1298        257        640:  51%|█████     | 409/809 [08:00<07:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.187     0.4295     0.1298        311        640:  51%|█████     | 410/809 [08:01<07:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.188     0.4294     0.1298        229        640:  51%|█████     | 412/809 [08:04<07:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.188     0.4294     0.1296        234        640:  51%|█████     | 414/809 [08:06<07:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.188     0.4295     0.1297        226        640:  51%|█████▏    | 415/809 [08:07<07:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.186       0.43     0.1295        194        640:  52%|█████▏    | 417/809 [08:10<07:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.186     0.4299     0.1294        271        640:  52%|█████▏    | 419/809 [08:12<07:54,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.186       0.43     0.1293        241        640:  52%|█████▏    | 420/809 [08:13<07:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.186     0.4301     0.1292        227        640:  52%|█████▏    | 422/809 [08:16<07:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.185     0.4302     0.1291        192        640:  52%|█████▏    | 424/809 [08:18<07:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.185     0.4302      0.129        245        640:  53%|█████▎    | 425/809 [08:19<07:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.184     0.4308     0.1289        121        640:  53%|█████▎    | 427/809 [08:22<07:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.185     0.4306     0.1288        342        640:  53%|█████▎    | 429/809 [08:24<07:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.186     0.4305     0.1289        638        640:  53%|█████▎    | 431/809 [08:26<07:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.186     0.4307     0.1289        328        640:  53%|█████▎    | 432/809 [08:27<07:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.184     0.4314     0.1287        159        640:  54%|█████▎    | 434/809 [08:30<07:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.184     0.4314     0.1289        142        640:  54%|█████▍    | 436/809 [08:32<07:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.183     0.4317     0.1289        221        640:  54%|█████▍    | 437/809 [08:33<07:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4321     0.1288        387        640:  54%|█████▍    | 439/809 [08:36<07:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182      0.432     0.1289        279        640:  55%|█████▍    | 441/809 [08:38<07:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4325     0.1287        141        640:  55%|█████▍    | 443/809 [08:40<07:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4326     0.1286        332        640:  55%|█████▍    | 444/809 [08:41<07:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.179     0.4326     0.1285        296        640:  55%|█████▌    | 446/809 [08:44<07:37,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.179     0.4329     0.1285        192        640:  55%|█████▌    | 447/809 [08:45<07:32,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.179     0.4329     0.1284        254        640:  56%|█████▌    | 449/809 [08:48<07:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.179     0.4329     0.1286        274        640:  56%|█████▌    | 451/809 [08:50<07:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4325     0.1289        485        640:  56%|█████▌    | 453/809 [08:52<07:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4323     0.1289        362        640:  56%|█████▌    | 454/809 [08:54<06:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4323      0.129        291        640:  56%|█████▋    | 456/809 [08:56<06:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4325      0.129        174        640:  57%|█████▋    | 458/809 [08:58<06:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4326     0.1289        114        640:  57%|█████▋    | 459/809 [08:59<06:43,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4323     0.1289        257        640:  57%|█████▋    | 461/809 [09:02<06:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4321     0.1289        290        640:  57%|█████▋    | 463/809 [09:04<06:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.183     0.4319     0.1289        350        640:  57%|█████▋    | 465/809 [09:06<06:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.183      0.432      0.129        203        640:  58%|█████▊    | 466/809 [09:07<06:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.183      0.432     0.1289        253        640:  58%|█████▊    | 468/809 [09:10<06:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4322     0.1287         51        640:  58%|█████▊    | 470/809 [09:12<06:27,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4321     0.1287        349        640:  58%|█████▊    | 472/809 [09:14<06:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4323     0.1286        185        640:  58%|█████▊    | 473/809 [09:16<06:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4324     0.1285        202        640:  59%|█████▊    | 475/809 [09:18<06:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4327     0.1283         90        640:  59%|█████▉    | 477/809 [09:20<06:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4328     0.1283        168        640:  59%|█████▉    | 478/809 [09:22<06:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4328     0.1282        210        640:  59%|█████▉    | 480/809 [09:24<06:47,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4326     0.1282        316        640:  60%|█████▉    | 482/809 [09:26<06:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4326     0.1281        169        640:  60%|█████▉    | 483/809 [09:28<06:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4324     0.1281        287        640:  60%|█████▉    | 485/809 [09:30<06:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4324     0.1282        396        640:  60%|██████    | 487/809 [09:32<06:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4324     0.1281        305        640:  60%|██████    | 488/809 [09:33<06:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4326      0.128        259        640:  61%|██████    | 490/809 [09:36<06:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4327     0.1279        254        640:  61%|██████    | 492/809 [09:38<06:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4324     0.1281        326        640:  61%|██████    | 494/809 [09:40<06:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4325      0.128        213        640:  61%|██████    | 495/809 [09:42<06:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4325      0.128        325        640:  61%|██████▏   | 497/809 [09:44<06:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4322     0.1283        479        640:  62%|██████▏   | 499/809 [09:46<06:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4321     0.1282        368        640:  62%|██████▏   | 501/809 [09:49<06:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4319     0.1283        363        640:  62%|██████▏   | 503/809 [09:51<05:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4319     0.1282        324        640:  62%|██████▏   | 504/809 [09:52<06:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4321     0.1281        393        640:  63%|██████▎   | 506/809 [09:55<05:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182      0.432     0.1281        268        640:  63%|██████▎   | 508/809 [09:57<05:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4321      0.128        260        640:  63%|██████▎   | 509/809 [09:58<06:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4325     0.1278        176        640:  63%|██████▎   | 511/809 [10:01<05:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4323      0.128        539        640:  63%|██████▎   | 513/809 [10:03<05:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4324     0.1279        226        640:  64%|██████▎   | 514/809 [10:04<05:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4326     0.1278        169        640:  64%|██████▍   | 516/809 [10:07<05:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4326     0.1278        185        640:  64%|██████▍   | 518/809 [10:09<05:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4327     0.1278        192        640:  64%|██████▍   | 519/809 [10:10<05:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.182     0.4328     0.1277        329        640:  64%|██████▍   | 521/809 [10:12<05:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4329     0.1277        232        640:  65%|██████▍   | 523/809 [10:15<05:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.181     0.4331     0.1276        239        640:  65%|██████▍   | 525/809 [10:17<05:27,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4332     0.1275        182        640:  65%|██████▌   | 526/809 [10:18<05:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4333     0.1274        239        640:  65%|██████▌   | 528/809 [10:21<05:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4334     0.1274        283        640:  66%|██████▌   | 530/809 [10:23<05:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G       1.18     0.4336     0.1273        250        640:  66%|██████▌   | 531/809 [10:24<05:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.179     0.4339     0.1272        333        640:  66%|██████▌   | 533/809 [10:26<05:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.179     0.4339     0.1272        223        640:  66%|██████▌   | 535/809 [10:29<05:14,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178      0.434     0.1272        124        640:  66%|██████▋   | 537/809 [10:31<05:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.179      0.434     0.1273        322        640:  67%|██████▋   | 538/809 [10:32<05:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4342     0.1272        248        640:  67%|██████▋   | 540/809 [10:35<05:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4343     0.1271        131        640:  67%|██████▋   | 542/809 [10:37<05:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4345      0.127        246        640:  67%|██████▋   | 544/809 [10:39<05:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4344      0.127        399        640:  67%|██████▋   | 545/809 [10:40<05:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4343     0.1271        258        640:  68%|██████▊   | 547/809 [10:43<05:00,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4343     0.1271        267        640:  68%|██████▊   | 549/809 [10:45<05:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4345      0.127        113        640:  68%|██████▊   | 550/809 [10:46<05:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4343     0.1271        264        640:  68%|██████▊   | 552/809 [10:49<05:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4344      0.127        192        640:  68%|██████▊   | 554/809 [10:51<05:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4344      0.127        157        640:  69%|██████▊   | 556/809 [10:53<05:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4345     0.1269        255        640:  69%|██████▉   | 557/809 [10:54<04:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4344      0.127        139        640:  69%|██████▉   | 559/809 [10:57<04:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.179     0.4343     0.1269        172        640:  69%|██████▉   | 561/809 [10:59<04:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4346     0.1269        148        640:  69%|██████▉   | 562/809 [11:00<04:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4346     0.1268        174        640:  70%|██████▉   | 564/809 [11:03<04:42,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4347     0.1266        232        640:  70%|██████▉   | 566/809 [11:05<04:39,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4349     0.1266        141        640:  70%|███████   | 567/809 [11:06<04:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4348     0.1267        151        640:  70%|███████   | 569/809 [11:09<04:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4352     0.1265        191        640:  71%|███████   | 571/809 [11:11<04:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4352     0.1265        144        640:  71%|███████   | 573/809 [11:13<04:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4353     0.1265         96        640:  71%|███████   | 574/809 [11:14<04:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4352     0.1264        260        640:  71%|███████   | 576/809 [11:17<04:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4351     0.1265        320        640:  71%|███████▏  | 578/809 [11:19<04:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4351     0.1266        217        640:  72%|███████▏  | 580/809 [11:21<04:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4352     0.1266        138        640:  72%|███████▏  | 581/809 [11:22<04:22,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4353     0.1265        218        640:  72%|███████▏  | 583/809 [11:25<04:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4355     0.1264        140        640:  72%|███████▏  | 585/809 [11:27<04:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4358     0.1264        314        640:  73%|███████▎  | 587/809 [11:29<04:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4357     0.1263        269        640:  73%|███████▎  | 588/809 [11:31<04:14,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4357     0.1264        151        640:  73%|███████▎  | 590/809 [11:33<04:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4356     0.1263        203        640:  73%|███████▎  | 592/809 [11:35<04:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4358     0.1262        274        640:  73%|███████▎  | 593/809 [11:36<04:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176      0.436     0.1262        176        640:  74%|███████▎  | 595/809 [11:39<04:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176      0.436     0.1262        266        640:  74%|███████▍  | 597/809 [11:41<04:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4362     0.1261        230        640:  74%|███████▍  | 598/809 [11:42<04:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4365      0.126        175        640:  74%|███████▍  | 600/809 [11:45<04:15,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4362     0.1264        385        640:  74%|███████▍  | 602/809 [11:47<04:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4362     0.1265        252        640:  75%|███████▍  | 604/809 [11:49<03:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4362     0.1264        204        640:  75%|███████▍  | 605/809 [11:51<03:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4363     0.1263        193        640:  75%|███████▌  | 607/809 [11:53<03:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4364     0.1262        233        640:  75%|███████▌  | 609/809 [11:55<03:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4366     0.1261        202        640:  75%|███████▌  | 610/809 [11:56<03:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4366     0.1262        204        640:  76%|███████▌  | 612/809 [11:59<04:01,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4367     0.1262        265        640:  76%|███████▌  | 614/809 [12:01<03:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4366     0.1264        550        640:  76%|███████▌  | 615/809 [12:02<03:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4368     0.1262        107        640:  76%|███████▋  | 617/809 [12:05<03:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4366     0.1262        423        640:  77%|███████▋  | 619/809 [12:07<03:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4366     0.1262        423        640:  77%|███████▋  | 620/809 [12:08<03:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4365     0.1263         67        640:  77%|███████▋  | 622/809 [12:11<03:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4364     0.1263        298        640:  77%|███████▋  | 624/809 [12:13<03:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4365     0.1262        184        640:  77%|███████▋  | 625/809 [12:14<03:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4365     0.1261        247        640:  78%|███████▊  | 627/809 [12:17<03:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4367      0.126        162        640:  78%|███████▊  | 629/809 [12:19<03:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4367      0.126        232        640:  78%|███████▊  | 631/809 [12:21<03:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4368      0.126        108        640:  78%|███████▊  | 632/809 [12:23<03:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4367     0.1261        373        640:  78%|███████▊  | 634/809 [12:25<03:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4367      0.126        206        640:  79%|███████▊  | 636/809 [12:27<03:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176      0.437      0.126        168        640:  79%|███████▊  | 637/809 [12:28<03:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4369     0.1259        111        640:  79%|███████▉  | 639/809 [12:31<03:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175      0.437     0.1259        148        640:  79%|███████▉  | 641/809 [12:33<03:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4367     0.1259        247        640:  79%|███████▉  | 643/809 [12:35<03:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4364     0.1262        373        640:  80%|███████▉  | 644/809 [12:37<03:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4367     0.1261         81        640:  80%|███████▉  | 646/809 [12:39<03:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4363     0.1262        249        640:  80%|████████  | 648/809 [12:41<03:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4362     0.1263        231        640:  80%|████████  | 650/809 [12:44<03:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4363     0.1262        269        640:  81%|████████  | 652/809 [12:46<03:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4362     0.1262        319        640:  81%|████████  | 653/809 [12:47<03:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4362     0.1264        363        640:  81%|████████  | 655/809 [12:49<02:57,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4366     0.1262         89        640:  81%|████████  | 657/809 [12:52<02:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4365     0.1263        303        640:  81%|████████▏ | 658/809 [12:53<03:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4364     0.1263        357        640:  82%|████████▏ | 660/809 [12:56<02:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4365     0.1263        342        640:  82%|████████▏ | 662/809 [12:58<03:00,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4366     0.1262        136        640:  82%|████████▏ | 663/809 [12:59<02:57,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4366     0.1264        146        640:  82%|████████▏ | 665/809 [13:02<02:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4366     0.1264        368        640:  82%|████████▏ | 667/809 [13:04<02:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4365     0.1266        368        640:  83%|████████▎ | 668/809 [13:05<02:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.178     0.4364     0.1266        259        640:  83%|████████▎ | 670/809 [13:07<02:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.177     0.4367     0.1265        138        640:  83%|████████▎ | 672/809 [13:10<02:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4368     0.1264        172        640:  83%|████████▎ | 674/809 [13:12<02:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4368     0.1264        313        640:  83%|████████▎ | 675/809 [13:13<02:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4368     0.1265        149        640:  84%|████████▎ | 677/809 [13:15<02:31,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4369     0.1264        189        640:  84%|████████▍ | 679/809 [13:18<02:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4371     0.1263        152        640:  84%|████████▍ | 681/809 [13:20<02:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4373     0.1263         87        640:  84%|████████▍ | 682/809 [13:21<02:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4374     0.1262        143        640:  85%|████████▍ | 684/809 [13:24<02:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4372     0.1264        191        640:  85%|████████▍ | 686/809 [13:26<02:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4371     0.1264        383        640:  85%|████████▍ | 687/809 [13:27<02:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4371     0.1264        313        640:  85%|████████▌ | 689/809 [13:29<02:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4372     0.1263        385        640:  85%|████████▌ | 691/809 [13:32<02:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4374     0.1263        213        640:  86%|████████▌ | 693/809 [13:34<02:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4373     0.1263        350        640:  86%|████████▌ | 694/809 [13:35<02:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4372     0.1263        224        640:  86%|████████▌ | 696/809 [13:38<02:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4371     0.1262        323        640:  86%|████████▋ | 698/809 [13:40<02:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4371     0.1262        197        640:  86%|████████▋ | 699/809 [13:41<02:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176      0.437     0.1262        413        640:  87%|████████▋ | 701/809 [13:44<02:04,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4369     0.1264        142        640:  87%|████████▋ | 703/809 [13:46<02:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4369     0.1264        345        640:  87%|████████▋ | 705/809 [13:48<02:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176      0.437     0.1263        295        640:  87%|████████▋ | 706/809 [13:49<01:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4371     0.1263        227        640:  88%|████████▊ | 708/809 [13:52<01:55,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4372     0.1262        295        640:  88%|████████▊ | 710/809 [13:54<01:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4372     0.1261        154        640:  88%|████████▊ | 711/809 [13:55<01:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4372     0.1262        620        640:  88%|████████▊ | 713/809 [13:58<01:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4371     0.1264        297        640:  88%|████████▊ | 715/809 [14:00<01:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4372     0.1263        305        640:  89%|████████▊ | 717/809 [14:02<01:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4372     0.1263        266        640:  89%|████████▉ | 718/809 [14:03<01:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4372     0.1263        282        640:  89%|████████▉ | 720/809 [14:06<01:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4375     0.1262        127        640:  89%|████████▉ | 722/809 [14:08<01:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4374     0.1263        421        640:  89%|████████▉ | 723/809 [14:09<01:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4372     0.1264        470        640:  90%|████████▉ | 725/809 [14:12<01:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4373     0.1263        225        640:  90%|████████▉ | 727/809 [14:14<01:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4373     0.1263        212        640:  90%|████████▉ | 728/809 [14:15<01:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4375     0.1262        180        640:  90%|█████████ | 730/809 [14:18<01:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4376     0.1261        173        640:  90%|█████████ | 732/809 [14:20<01:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4376     0.1261        196        640:  91%|█████████ | 733/809 [14:21<01:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173     0.4377     0.1259        290        640:  91%|█████████ | 735/809 [14:24<01:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4375      0.126        456        640:  91%|█████████ | 737/809 [14:26<01:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173     0.4376     0.1259        382        640:  91%|█████████▏| 739/809 [14:28<01:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173     0.4376     0.1259        190        640:  91%|█████████▏| 740/809 [14:29<01:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173     0.4376     0.1258        132        640:  92%|█████████▏| 742/809 [14:32<01:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.172     0.4377     0.1258        139        640:  92%|█████████▏| 743/809 [14:33<01:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4373     0.1263        318        640:  92%|█████████▏| 745/809 [14:36<01:18,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4373     0.1262        239        640:  92%|█████████▏| 747/809 [14:38<01:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4371     0.1262        233        640:  93%|█████████▎| 749/809 [14:40<01:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175      0.437     0.1265        425        640:  93%|█████████▎| 750/809 [14:41<01:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4372     0.1264        142        640:  93%|█████████▎| 752/809 [14:44<01:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4373     0.1264        226        640:  93%|█████████▎| 754/809 [14:46<01:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4373     0.1263        165        640:  93%|█████████▎| 755/809 [14:47<01:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4372     0.1263        174        640:  94%|█████████▎| 757/809 [14:50<01:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4371     0.1264        122        640:  94%|█████████▍| 759/809 [14:52<00:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175      0.437     0.1265        166        640:  94%|█████████▍| 761/809 [14:54<00:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4371     0.1264        183        640:  94%|█████████▍| 762/809 [14:56<00:58,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175      0.437     0.1264        228        640:  94%|█████████▍| 764/809 [14:58<00:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4369     0.1264        327        640:  95%|█████████▍| 765/809 [14:59<00:53,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176     0.4367     0.1263        243        640:  95%|█████████▍| 767/809 [15:02<00:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176      0.437     0.1263        309        640:  95%|█████████▌| 769/809 [15:04<00:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.176      0.437     0.1263        414        640:  95%|█████████▌| 771/809 [15:06<00:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4371     0.1262        274        640:  95%|█████████▌| 772/809 [15:08<00:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.175     0.4374     0.1261        389        640:  96%|█████████▌| 774/809 [15:10<00:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4376      0.126        130        640:  96%|█████████▌| 776/809 [15:12<00:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4378      0.126         78        640:  96%|█████████▌| 777/809 [15:13<00:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.174     0.4377     0.1259        243        640:  96%|█████████▋| 779/809 [15:16<00:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173      0.438     0.1258        412        640:  97%|█████████▋| 781/809 [15:18<00:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173     0.4381     0.1258        202        640:  97%|█████████▋| 783/809 [15:21<00:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173     0.4381     0.1258        371        640:  97%|█████████▋| 784/809 [15:22<00:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.173     0.4383     0.1257        211        640:  97%|█████████▋| 786/809 [15:24<00:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.172     0.4386     0.1256        232        640:  97%|█████████▋| 788/809 [15:27<00:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.172     0.4385     0.1256        453        640:  98%|█████████▊| 789/809 [15:28<00:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.172     0.4386     0.1255        458        640:  98%|█████████▊| 791/809 [15:30<00:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.172     0.4386     0.1254        350        640:  98%|█████████▊| 793/809 [15:32<00:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.172     0.4387     0.1253        168        640:  98%|█████████▊| 795/809 [15:35<00:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.172     0.4387     0.1253        282        640:  98%|█████████▊| 796/809 [15:36<00:14,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.172     0.4387     0.1253        267        640:  99%|█████████▊| 798/809 [15:38<00:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.172     0.4388     0.1253        141        640:  99%|█████████▉| 800/809 [15:40<00:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.171      0.439     0.1252        188        640:  99%|█████████▉| 802/809 [15:43<00:08,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.171     0.4393     0.1251        271        640:  99%|█████████▉| 804/809 [15:45<00:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.171     0.4393     0.1251        121        640: 100%|█████████▉| 805/809 [15:46<00:04,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     70/300      13.3G      1.171     0.4394      0.125        130        640: 100%|█████████▉| 807/809 [15:48<00:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   0%|          | 0/69 [00:00<?, ?it/s]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  13%|█▎        | 9/69 [00:02<00:14,  4.18it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  25%|██▍       | 17/69 [00:04<00:12,  4.07it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  36%|███▌      | 25/69 [00:06<00:10,  4.10it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  48%|████▊     | 33/69 [00:08<00:08,  4.04it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  59%|█████▉    | 41/69 [00:10<00:06,  4.07it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  71%|███████   | 49/69 [00:12<00:05,  3.77it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  83%|████████▎ | 57/69 [00:14<00:03,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:00,  4.10it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.7G       1.33     0.3798     0.1195        278        640:   0%|          | 1/809 [00:01<15:04,  1.12s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.7G      1.545      0.337     0.1797        245        640:   0%|          | 3/809 [00:03<15:08,  1.13s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.7G      1.432     0.3714     0.1547         86        640:   1%|          | 5/809 [00:05<15:20,  1.14s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.7G      1.383      0.385     0.1432        107        640:   1%|          | 6/809 [00:06<15:21,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.371     0.3866      0.132        287        640:   1%|          | 8/809 [00:09<15:27,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.345     0.3943     0.1345        222        640:   1%|          | 10/809 [00:11<15:19,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.285     0.4175     0.1292        208        640:   1%|▏         | 12/809 [00:13<15:13,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.267     0.4218     0.1275        250        640:   2%|▏         | 13/809 [00:14<15:09,  1.14s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.264     0.4216     0.1296        329        640:   2%|▏         | 15/809 [00:17<15:33,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.233     0.4295     0.1239        334        640:   2%|▏         | 17/809 [00:19<15:27,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.24     0.4289     0.1257        401        640:   2%|▏         | 19/809 [00:21<15:17,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.227     0.4349     0.1236        180        640:   2%|▏         | 20/809 [00:23<15:16,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.245     0.4325     0.1319        438        640:   3%|▎         | 22/809 [00:25<15:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.229      0.437     0.1284        150        640:   3%|▎         | 24/809 [00:27<15:21,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.233     0.4349     0.1279        335        640:   3%|▎         | 25/809 [00:29<15:19,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.228     0.4374     0.1273        140        640:   3%|▎         | 27/809 [00:31<15:35,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.218     0.4389     0.1253        207        640:   4%|▎         | 29/809 [00:33<15:14,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.214     0.4392     0.1238        195        640:   4%|▎         | 30/809 [00:34<15:04,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.205     0.4422     0.1227        237        640:   4%|▍         | 32/809 [00:37<14:59,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.198     0.4441     0.1213        360        640:   4%|▍         | 34/809 [00:39<15:16,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.189     0.4481     0.1211        245        640:   4%|▍         | 36/809 [00:42<15:12,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.202     0.4456     0.1299        444        640:   5%|▍         | 37/809 [00:43<15:20,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.193     0.4479     0.1278        220        640:   5%|▍         | 39/809 [00:45<15:41,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.193     0.4488     0.1273        168        640:   5%|▌         | 41/809 [00:47<15:10,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.183     0.4522     0.1262        145        640:   5%|▌         | 42/809 [00:49<15:00,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.168     0.4555      0.124        138        640:   5%|▌         | 44/809 [00:51<14:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.171     0.4543     0.1246        415        640:   6%|▌         | 46/809 [00:53<14:54,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.178     0.4526     0.1257        430        640:   6%|▌         | 47/809 [00:55<15:06,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.189     0.4479     0.1272        379        640:   6%|▌         | 49/809 [00:57<14:52,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.176     0.4521     0.1259        187        640:   6%|▋         | 51/809 [00:59<14:38,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.168     0.4539     0.1253        153        640:   7%|▋         | 53/809 [01:02<14:47,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.161     0.4555     0.1247        213        640:   7%|▋         | 54/809 [01:03<14:51,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.164     0.4549     0.1255        382        640:   7%|▋         | 56/809 [01:05<14:41,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.169     0.4526     0.1252        421        640:   7%|▋         | 58/809 [01:08<15:00,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.165     0.4534     0.1245        253        640:   7%|▋         | 59/809 [01:09<14:44,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.172     0.4511     0.1254        438        640:   8%|▊         | 61/809 [01:11<14:31,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.174     0.4503     0.1249        235        640:   8%|▊         | 63/809 [01:13<14:32,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.178     0.4491     0.1262        662        640:   8%|▊         | 64/809 [01:14<14:27,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.179      0.449     0.1282        331        640:   8%|▊         | 66/809 [01:17<14:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.173     0.4516     0.1272        312        640:   8%|▊         | 68/809 [01:19<14:52,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.164     0.4532      0.126        259        640:   9%|▊         | 70/809 [01:22<14:32,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.158     0.4541     0.1253        200        640:   9%|▉         | 71/809 [01:23<14:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.157      0.455     0.1249        262        640:   9%|▉         | 73/809 [01:25<14:33,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.153      0.457     0.1245        203        640:   9%|▉         | 75/809 [01:28<14:15,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.155     0.4564     0.1248        442        640:   9%|▉         | 76/809 [01:29<14:10,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.159     0.4556     0.1282        141        640:  10%|▉         | 78/809 [01:31<14:37,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.155     0.4568     0.1275        292        640:  10%|▉         | 80/809 [01:34<14:49,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.151     0.4585     0.1268         95        640:  10%|█         | 81/809 [01:35<14:35,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.155     0.4571     0.1278        323        640:  10%|█         | 83/809 [01:37<14:15,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.149     0.4589     0.1268        121        640:  11%|█         | 85/809 [01:39<14:08,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.147     0.4594     0.1263        237        640:  11%|█         | 86/809 [01:41<14:02,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.146     0.4593     0.1267        176        640:  11%|█         | 88/809 [01:43<14:26,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.148     0.4579     0.1264        311        640:  11%|█         | 90/809 [01:45<14:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.148     0.4574     0.1257        221        640:  11%|█▏        | 92/809 [01:48<13:51,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.146     0.4584     0.1252        179        640:  11%|█▏        | 93/809 [01:49<14:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.147     0.4581      0.125        219        640:  12%|█▏        | 95/809 [01:51<13:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.14     0.4598      0.124        216        640:  12%|█▏        | 97/809 [01:54<13:55,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.139       0.46     0.1239        312        640:  12%|█▏        | 98/809 [01:55<13:52,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.133     0.4614     0.1229        113        640:  12%|█▏        | 100/809 [01:57<13:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.137     0.4609     0.1254        498        640:  13%|█▎        | 102/809 [01:59<14:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.139     0.4597     0.1255        333        640:  13%|█▎        | 104/809 [02:02<13:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.137     0.4601     0.1251        235        640:  13%|█▎        | 105/809 [02:03<13:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.138       0.46     0.1248        240        640:  13%|█▎        | 107/809 [02:05<13:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.145     0.4579     0.1251        419        640:  13%|█▎        | 109/809 [02:08<13:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.146     0.4574      0.125        216        640:  14%|█▎        | 110/809 [02:09<13:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.144     0.4577     0.1244        242        640:  14%|█▍        | 112/809 [02:11<13:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.146     0.4572     0.1242        273        640:  14%|█▍        | 114/809 [02:14<13:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.146     0.4573     0.1244        269        640:  14%|█▍        | 116/809 [02:16<13:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.146     0.4566     0.1242        341        640:  14%|█▍        | 117/809 [02:17<13:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.145     0.4563     0.1239        224        640:  15%|█▍        | 119/809 [02:19<13:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.148     0.4557     0.1241        345        640:  15%|█▍        | 121/809 [02:22<13:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.149     0.4555     0.1243        241        640:  15%|█▌        | 122/809 [02:23<13:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.158     0.4527     0.1263        489        640:  15%|█▌        | 124/809 [02:25<13:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.159     0.4523     0.1259        335        640:  16%|█▌        | 126/809 [02:28<13:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.164     0.4511     0.1272        357        640:  16%|█▌        | 127/809 [02:29<13:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.162     0.4511     0.1267        465        640:  16%|█▌        | 129/809 [02:31<13:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.157     0.4521     0.1263        312        640:  16%|█▌        | 131/809 [02:33<13:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.152     0.4536     0.1259        191        640:  16%|█▋        | 133/809 [02:36<13:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.15     0.4544     0.1255        220        640:  17%|█▋        | 135/809 [02:38<13:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.15     0.4541     0.1253        303        640:  17%|█▋        | 136/809 [02:39<13:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.152     0.4542      0.126        210        640:  17%|█▋        | 138/809 [02:42<13:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.153      0.454     0.1259        146        640:  17%|█▋        | 140/809 [02:44<13:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.152     0.4541     0.1255        156        640:  18%|█▊        | 142/809 [02:46<13:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.153     0.4538     0.1255        330        640:  18%|█▊        | 143/809 [02:48<13:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.152     0.4542      0.125        346        640:  18%|█▊        | 145/809 [02:50<13:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.151     0.4547     0.1247        239        640:  18%|█▊        | 147/809 [02:52<13:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.151      0.455     0.1246        161        640:  18%|█▊        | 148/809 [02:54<13:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.153     0.4542     0.1249        251        640:  19%|█▊        | 150/809 [02:56<12:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.151     0.4552     0.1244        198        640:  19%|█▉        | 152/809 [02:58<13:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.152      0.455     0.1245        307        640:  19%|█▉        | 153/809 [03:00<12:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.154     0.4545     0.1246        254        640:  19%|█▉        | 155/809 [03:02<12:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.155      0.454     0.1247        268        640:  19%|█▉        | 157/809 [03:04<12:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.156     0.4535     0.1246        344        640:  20%|█▉        | 158/809 [03:05<13:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.157     0.4535     0.1245        235        640:  20%|█▉        | 160/809 [03:08<12:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.157     0.4531     0.1246         89        640:  20%|██        | 162/809 [03:10<12:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.157     0.4532     0.1247        278        640:  20%|██        | 163/809 [03:11<12:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.152     0.4547     0.1242        121        640:  20%|██        | 165/809 [03:14<12:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.149     0.4558     0.1237        192        640:  21%|██        | 167/809 [03:16<12:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.149     0.4557     0.1239        421        640:  21%|██        | 169/809 [03:18<12:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.147     0.4566     0.1236        134        640:  21%|██        | 170/809 [03:20<12:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.145     0.4574     0.1235        193        640:  21%|██▏       | 172/809 [03:22<12:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.143     0.4579     0.1231        203        640:  22%|██▏       | 174/809 [03:24<12:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.14     0.4587     0.1229        177        640:  22%|██▏       | 175/809 [03:26<12:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.137     0.4598     0.1225        159        640:  22%|██▏       | 177/809 [03:28<12:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.138     0.4592     0.1224        203        640:  22%|██▏       | 179/809 [03:30<12:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.138     0.4591     0.1222        339        640:  22%|██▏       | 181/809 [03:33<12:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.135     0.4603     0.1221        105        640:  22%|██▏       | 182/809 [03:34<12:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.131     0.4614     0.1217        145        640:  23%|██▎       | 184/809 [03:36<12:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.13     0.4615     0.1214        213        640:  23%|██▎       | 186/809 [03:38<12:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.13     0.4616     0.1213        102        640:  23%|██▎       | 187/809 [03:40<12:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.129      0.462     0.1211        253        640:  23%|██▎       | 189/809 [03:42<12:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.129     0.4617     0.1211        470        640:  24%|██▎       | 191/809 [03:44<12:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.129     0.4619      0.121        289        640:  24%|██▎       | 192/809 [03:46<12:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.129     0.4616     0.1214        349        640:  24%|██▍       | 194/809 [03:48<12:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.124     0.4631     0.1208        108        640:  24%|██▍       | 196/809 [03:50<12:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.124     0.4635     0.1208        155        640:  24%|██▍       | 197/809 [03:52<12:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.124     0.4635     0.1213        259        640:  25%|██▍       | 199/809 [03:54<12:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.124     0.4634     0.1214        361        640:  25%|██▍       | 201/809 [03:56<12:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.124     0.4633     0.1212        257        640:  25%|██▌       | 203/809 [03:59<11:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.125     0.4629     0.1214        419        640:  25%|██▌       | 204/809 [04:00<11:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.124     0.4629     0.1213        301        640:  25%|██▌       | 206/809 [04:02<11:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.124      0.463     0.1215        358        640:  26%|██▌       | 208/809 [04:05<11:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.122     0.4636     0.1214        156        640:  26%|██▌       | 209/809 [04:06<11:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.122     0.4638     0.1215        170        640:  26%|██▌       | 211/809 [04:08<11:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.121     0.4646     0.1212        221        640:  26%|██▋       | 213/809 [04:10<11:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.119     0.4647      0.121        237        640:  27%|██▋       | 215/809 [04:13<11:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.119     0.4649      0.121        360        640:  27%|██▋       | 216/809 [04:14<11:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4655     0.1208        267        640:  27%|██▋       | 218/809 [04:16<11:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.114     0.4662     0.1204        288        640:  27%|██▋       | 220/809 [04:19<11:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.113     0.4663     0.1205        246        640:  27%|██▋       | 221/809 [04:20<11:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.111     0.4669     0.1203         84        640:  28%|██▊       | 223/809 [04:22<11:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.111      0.467     0.1203        327        640:  28%|██▊       | 225/809 [04:24<11:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4667     0.1205        530        640:  28%|██▊       | 226/809 [04:26<11:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.111      0.467     0.1204        267        640:  28%|██▊       | 228/809 [04:28<11:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.113     0.4662     0.1205        244        640:  28%|██▊       | 230/809 [04:30<11:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4662     0.1204        293        640:  29%|██▊       | 232/809 [04:33<11:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.11     0.4665     0.1203        150        640:  29%|██▉       | 233/809 [04:34<11:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4667     0.1201        155        640:  29%|██▉       | 235/809 [04:36<11:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.108     0.4668     0.1198        224        640:  29%|██▉       | 237/809 [04:39<11:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4664     0.1203        456        640:  29%|██▉       | 238/809 [04:40<11:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4666     0.1201        277        640:  30%|██▉       | 240/809 [04:42<11:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.11     0.4659     0.1206        187        640:  30%|██▉       | 242/809 [04:45<11:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4655     0.1211        372        640:  30%|███       | 243/809 [04:46<11:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.115     0.4646     0.1216        509        640:  30%|███       | 245/809 [04:48<11:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.115     0.4645     0.1217        305        640:  31%|███       | 247/809 [04:50<11:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4642     0.1222        340        640:  31%|███       | 249/809 [04:53<10:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4643     0.1221        286        640:  31%|███       | 250/809 [04:54<10:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4644     0.1224        181        640:  31%|███       | 252/809 [04:56<10:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.118     0.4638     0.1228        355        640:  31%|███▏      | 254/809 [04:59<10:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.119     0.4635     0.1231        285        640:  32%|███▏      | 255/809 [05:00<10:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.12     0.4632      0.123        252        640:  32%|███▏      | 257/809 [05:02<10:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.12     0.4634     0.1228        184        640:  32%|███▏      | 259/809 [05:04<10:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.121     0.4632     0.1232        273        640:  32%|███▏      | 261/809 [05:07<10:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.119     0.4635     0.1231        166        640:  32%|███▏      | 262/809 [05:08<10:28,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.118     0.4638     0.1229        258        640:  33%|███▎      | 264/809 [05:10<10:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.117     0.4639     0.1229        186        640:  33%|███▎      | 266/809 [05:12<10:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4641     0.1229        293        640:  33%|███▎      | 268/809 [05:15<10:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.115     0.4641     0.1228        188        640:  33%|███▎      | 269/809 [05:16<10:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.114     0.4642     0.1226        193        640:  33%|███▎      | 271/809 [05:18<10:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4647     0.1223        191        640:  34%|███▎      | 273/809 [05:21<10:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.113     0.4645     0.1224        365        640:  34%|███▍      | 274/809 [05:22<10:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4645     0.1223        246        640:  34%|███▍      | 276/809 [05:24<10:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.113     0.4643     0.1222        154        640:  34%|███▍      | 278/809 [05:26<10:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.111     0.4645     0.1219        201        640:  35%|███▍      | 280/809 [05:29<10:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.111     0.4646     0.1219        362        640:  35%|███▍      | 282/809 [05:31<10:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.11     0.4647     0.1218        180        640:  35%|███▍      | 283/809 [05:32<10:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112      0.464      0.122        487        640:  35%|███▌      | 285/809 [05:35<10:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4631     0.1224        458        640:  35%|███▌      | 287/809 [05:37<10:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4628     0.1223        399        640:  36%|███▌      | 289/809 [05:39<10:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.117     0.4623     0.1224        480        640:  36%|███▌      | 290/809 [05:41<10:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.117     0.4623     0.1225        171        640:  36%|███▌      | 292/809 [05:43<10:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.118     0.4618     0.1232        526        640:  36%|███▋      | 294/809 [05:45<10:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.119     0.4616     0.1232        440        640:  36%|███▋      | 295/809 [05:46<10:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.118     0.4619      0.123        180        640:  37%|███▋      | 297/809 [05:49<10:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.117     0.4619     0.1229        460        640:  37%|███▋      | 299/809 [05:51<10:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4623     0.1228        120        640:  37%|███▋      | 300/809 [05:52<10:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4622     0.1236        230        640:  37%|███▋      | 302/809 [05:55<10:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4625     0.1235        350        640:  38%|███▊      | 304/809 [05:57<09:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4625     0.1233        382        640:  38%|███▊      | 306/809 [06:00<09:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.116     0.4624     0.1232        272        640:  38%|███▊      | 307/809 [06:01<09:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.115     0.4627      0.123        324        640:  38%|███▊      | 309/809 [06:03<10:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.115     0.4631     0.1231         71        640:  38%|███▊      | 311/809 [06:06<09:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.115     0.4633      0.123        270        640:  39%|███▊      | 312/809 [06:07<09:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.115     0.4638      0.123        163        640:  39%|███▉      | 314/809 [06:09<09:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.114     0.4637     0.1229        348        640:  39%|███▉      | 316/809 [06:11<09:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.114     0.4639     0.1229        277        640:  39%|███▉      | 317/809 [06:12<09:27,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4643     0.1226        270        640:  39%|███▉      | 319/809 [06:15<09:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.113      0.464     0.1229        177        640:  40%|███▉      | 321/809 [06:17<09:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.113     0.4642     0.1227        194        640:  40%|███▉      | 323/809 [06:19<09:17,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4645     0.1226        170        640:  40%|████      | 324/809 [06:21<09:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4651     0.1223        212        640:  40%|████      | 326/809 [06:23<09:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4652     0.1222        149        640:  41%|████      | 328/809 [06:25<09:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.108     0.4653     0.1221        170        640:  41%|████      | 329/809 [06:27<09:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.107     0.4655     0.1219        151        640:  41%|████      | 331/809 [06:29<09:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106     0.4658     0.1218        208        640:  41%|████      | 333/809 [06:31<09:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106     0.4654     0.1217        280        640:  41%|████▏     | 335/809 [06:34<09:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106     0.4656     0.1216        269        640:  42%|████▏     | 336/809 [06:35<09:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106     0.4655     0.1215        297        640:  42%|████▏     | 338/809 [06:37<09:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.107      0.465     0.1219        215        640:  42%|████▏     | 340/809 [06:39<09:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.108     0.4646     0.1218        276        640:  42%|████▏     | 342/809 [06:42<08:57,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.108     0.4644     0.1217        180        640:  42%|████▏     | 343/809 [06:43<09:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.11     0.4641      0.122        379        640:  43%|████▎     | 345/809 [06:45<08:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.111     0.4637     0.1222        366        640:  43%|████▎     | 347/809 [06:47<08:53,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4635     0.1222        252        640:  43%|████▎     | 348/809 [06:49<08:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.11     0.4638     0.1221         97        640:  43%|████▎     | 350/809 [06:51<08:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.111     0.4638     0.1221        225        640:  44%|████▎     | 352/809 [06:53<09:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4636     0.1225        229        640:  44%|████▍     | 354/809 [06:56<08:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4635     0.1225        121        640:  44%|████▍     | 355/809 [06:57<08:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.113     0.4632     0.1225        176        640:  44%|████▍     | 357/809 [06:59<08:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.115     0.4624     0.1229        438        640:  44%|████▍     | 359/809 [07:02<08:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.114     0.4629     0.1228        259        640:  44%|████▍     | 360/809 [07:03<08:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.113     0.4634     0.1227         50        640:  45%|████▍     | 362/809 [07:05<08:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.113     0.4634     0.1226        333        640:  45%|████▍     | 364/809 [07:07<08:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4634     0.1225        204        640:  45%|████▌     | 366/809 [07:10<08:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4634     0.1225        155        640:  45%|████▌     | 367/809 [07:11<08:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.113     0.4631     0.1227        287        640:  46%|████▌     | 369/809 [07:13<08:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4633     0.1226        173        640:  46%|████▌     | 371/809 [07:15<08:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.113      0.463     0.1229        639        640:  46%|████▌     | 372/809 [07:17<08:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.112     0.4635     0.1227        189        640:  46%|████▌     | 374/809 [07:19<08:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G       1.11     0.4639     0.1225        271        640:  46%|████▋     | 376/809 [07:21<08:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4643     0.1222        200        640:  47%|████▋     | 378/809 [07:24<08:17,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.108     0.4646     0.1221        100        640:  47%|████▋     | 379/809 [07:25<08:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.108     0.4645      0.122        286        640:  47%|████▋     | 381/809 [07:27<08:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.107     0.4647      0.122        351        640:  47%|████▋     | 383/809 [07:30<08:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.107     0.4646     0.1219        209        640:  47%|████▋     | 384/809 [07:31<08:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.108     0.4642     0.1219        463        640:  48%|████▊     | 386/809 [07:33<08:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109      0.464     0.1218        216        640:  48%|████▊     | 388/809 [07:35<08:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4638     0.1218        199        640:  48%|████▊     | 390/809 [07:38<08:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4637     0.1219        210        640:  48%|████▊     | 391/809 [07:39<08:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4637     0.1217        225        640:  49%|████▊     | 393/809 [07:41<08:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4636     0.1217        342        640:  49%|████▉     | 395/809 [07:44<08:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4638     0.1216        245        640:  49%|████▉     | 396/809 [07:45<08:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4637     0.1216        176        640:  49%|████▉     | 398/809 [07:47<08:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4635     0.1214        189        640:  49%|████▉     | 400/809 [07:50<08:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.108     0.4638     0.1213        305        640:  50%|████▉     | 402/809 [07:52<07:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4637     0.1212        364        640:  50%|████▉     | 403/809 [07:53<07:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.108     0.4639     0.1211        162        640:  50%|█████     | 405/809 [07:55<07:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.109     0.4637      0.121        336        640:  50%|█████     | 407/809 [07:58<07:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.108      0.464     0.1209        140        640:  50%|█████     | 408/809 [07:59<07:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106     0.4644     0.1207        106        640:  51%|█████     | 410/809 [08:01<07:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4649     0.1204        223        640:  51%|█████     | 412/809 [08:03<07:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4652     0.1203        242        640:  51%|█████     | 414/809 [08:06<07:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4653     0.1203        219        640:  51%|█████▏    | 415/809 [08:07<07:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103     0.4653     0.1201        252        640:  52%|█████▏    | 417/809 [08:09<07:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103     0.4657       0.12        206        640:  52%|█████▏    | 419/809 [08:12<07:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4659     0.1198        144        640:  52%|█████▏    | 420/809 [08:13<07:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103     0.4657     0.1198        329        640:  52%|█████▏    | 422/809 [08:15<07:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102      0.466     0.1198        162        640:  52%|█████▏    | 424/809 [08:18<07:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4663     0.1197        127        640:  53%|█████▎    | 425/809 [08:19<07:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4661     0.1198        518        640:  53%|█████▎    | 427/809 [08:21<07:20,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103      0.466     0.1199        311        640:  53%|█████▎    | 429/809 [08:23<07:18,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4663     0.1197        133        640:  53%|█████▎    | 431/809 [08:26<07:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4665     0.1195        134        640:  54%|█████▎    | 433/809 [08:28<07:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4665     0.1194        170        640:  54%|█████▎    | 434/809 [08:29<07:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4666     0.1194        540        640:  54%|█████▍    | 436/809 [08:32<07:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4671     0.1193        201        640:  54%|█████▍    | 438/809 [08:34<07:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1      0.467     0.1194        393        640:  54%|█████▍    | 439/809 [08:35<07:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4671     0.1192        222        640:  55%|█████▍    | 441/809 [08:38<07:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4671     0.1193        304        640:  55%|█████▍    | 443/809 [08:40<07:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4674      0.119        192        640:  55%|█████▌    | 445/809 [08:42<07:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4675      0.119        244        640:  55%|█████▌    | 446/809 [08:43<07:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4674      0.119        286        640:  55%|█████▌    | 448/809 [08:46<07:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4671     0.1189        129        640:  56%|█████▌    | 450/809 [08:48<06:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4673     0.1188        228        640:  56%|█████▌    | 452/809 [08:50<06:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4673     0.1187        192        640:  56%|█████▌    | 453/809 [08:52<06:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4673     0.1186        205        640:  56%|█████▌    | 455/809 [08:54<06:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4675     0.1184        105        640:  56%|█████▋    | 457/809 [08:56<06:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4677     0.1183        159        640:  57%|█████▋    | 458/809 [08:57<06:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4678     0.1183         86        640:  57%|█████▋    | 460/809 [09:00<06:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4677     0.1183        208        640:  57%|█████▋    | 462/809 [09:02<06:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.096      0.468     0.1181        113        640:  57%|█████▋    | 464/809 [09:04<06:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.096      0.468     0.1181        177        640:  57%|█████▋    | 465/809 [09:06<06:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.096     0.4681     0.1179        240        640:  58%|█████▊    | 467/809 [09:08<06:34,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4678     0.1179        225        640:  58%|█████▊    | 469/809 [09:10<06:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4678     0.1179        219        640:  58%|█████▊    | 470/809 [09:11<06:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4675     0.1179        170        640:  58%|█████▊    | 472/809 [09:14<06:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4675     0.1178        118        640:  59%|█████▊    | 474/809 [09:16<06:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4671     0.1183        321        640:  59%|█████▉    | 476/809 [09:19<06:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4669     0.1184        587        640:  59%|█████▉    | 477/809 [09:20<06:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4668     0.1185        377        640:  59%|█████▉    | 479/809 [09:22<06:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4667     0.1185        419        640:  59%|█████▉    | 481/809 [09:24<06:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4665     0.1185        387        640:  60%|█████▉    | 482/809 [09:26<06:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4665     0.1184        271        640:  60%|█████▉    | 484/809 [09:28<06:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103     0.4663     0.1183        267        640:  60%|██████    | 486/809 [09:30<06:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4661     0.1186        408        640:  60%|██████    | 487/809 [09:32<06:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4661     0.1187        196        640:  60%|██████    | 489/809 [09:34<06:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4659     0.1189        264        640:  61%|██████    | 491/809 [09:36<06:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105      0.466      0.119        434        640:  61%|██████    | 492/809 [09:38<06:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4659     0.1191        133        640:  61%|██████    | 494/809 [09:40<06:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104      0.466      0.119        267        640:  61%|██████▏   | 496/809 [09:42<06:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4658      0.119        303        640:  62%|██████▏   | 498/809 [09:45<06:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4656      0.119        264        640:  62%|██████▏   | 499/809 [09:46<06:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4659      0.119        106        640:  62%|██████▏   | 501/809 [09:48<06:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103      0.466     0.1188        133        640:  62%|██████▏   | 503/809 [09:51<06:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103      0.466     0.1187        176        640:  62%|██████▏   | 504/809 [09:52<06:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4662     0.1186        149        640:  63%|██████▎   | 506/809 [09:54<05:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4663     0.1185        269        640:  63%|██████▎   | 508/809 [09:56<05:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4662     0.1185        316        640:  63%|██████▎   | 509/809 [09:58<05:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4663     0.1183        162        640:  63%|██████▎   | 511/809 [10:00<05:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4661     0.1183        234        640:  63%|██████▎   | 513/809 [10:02<05:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4664     0.1182        312        640:  64%|██████▎   | 515/809 [10:05<05:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4664     0.1182        103        640:  64%|██████▍   | 516/809 [10:06<05:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4663     0.1181        217        640:  64%|██████▍   | 518/809 [10:08<05:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4664     0.1182        111        640:  64%|██████▍   | 520/809 [10:11<05:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4664     0.1181        199        640:  64%|██████▍   | 521/809 [10:12<05:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4665     0.1181        191        640:  65%|██████▍   | 523/809 [10:14<05:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4664     0.1181        240        640:  65%|██████▍   | 525/809 [10:16<05:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4665      0.118        232        640:  65%|██████▌   | 527/809 [10:19<05:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4666     0.1179         98        640:  65%|██████▌   | 528/809 [10:20<05:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4667     0.1178        149        640:  66%|██████▌   | 530/809 [10:22<05:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4666     0.1177        277        640:  66%|██████▌   | 532/809 [10:24<05:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4666     0.1177        253        640:  66%|██████▌   | 533/809 [10:26<05:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4667     0.1176        184        640:  66%|██████▌   | 535/809 [10:28<05:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4669     0.1174        194        640:  66%|██████▋   | 537/809 [10:30<05:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4669     0.1174        332        640:  67%|██████▋   | 539/809 [10:33<05:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4668     0.1174        204        640:  67%|██████▋   | 540/809 [10:34<05:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4669     0.1172        329        640:  67%|██████▋   | 542/809 [10:36<05:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4668     0.1173        324        640:  67%|██████▋   | 544/809 [10:39<05:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4668     0.1173        204        640:  67%|██████▋   | 545/809 [10:40<05:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4669     0.1172        323        640:  68%|██████▊   | 547/809 [10:42<05:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4671     0.1173        266        640:  68%|██████▊   | 549/809 [10:44<05:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4673     0.1172        170        640:  68%|██████▊   | 551/809 [10:47<05:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4673     0.1172        255        640:  68%|██████▊   | 552/809 [10:48<05:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098      0.467     0.1172        396        640:  68%|██████▊   | 554/809 [10:50<04:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4672     0.1171        227        640:  69%|██████▊   | 556/809 [10:52<04:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4672     0.1173        296        640:  69%|██████▉   | 557/809 [10:54<04:49,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4672     0.1178        171        640:  69%|██████▉   | 559/809 [10:56<04:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4674     0.1177        115        640:  69%|██████▉   | 561/809 [10:58<04:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4674     0.1176        169        640:  70%|██████▉   | 563/809 [11:01<04:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4672     0.1181        372        640:  70%|██████▉   | 564/809 [11:02<04:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4672     0.1181        185        640:  70%|██████▉   | 566/809 [11:04<04:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4669     0.1188        385        640:  70%|███████   | 568/809 [11:07<04:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099      0.467     0.1187        137        640:  70%|███████   | 570/809 [11:09<04:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099      0.467     0.1186        178        640:  71%|███████   | 571/809 [11:10<04:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4671     0.1185        122        640:  71%|███████   | 573/809 [11:12<04:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4673     0.1184        316        640:  71%|███████   | 575/809 [11:15<04:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4673     0.1184        171        640:  71%|███████   | 576/809 [11:16<04:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4673     0.1183        117        640:  71%|███████▏  | 578/809 [11:18<04:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.096     0.4674     0.1182        318        640:  72%|███████▏  | 580/809 [11:21<04:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.095     0.4676     0.1181         99        640:  72%|███████▏  | 581/809 [11:22<04:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.096     0.4673     0.1182        200        640:  72%|███████▏  | 583/809 [11:24<04:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4669     0.1183        229        640:  72%|███████▏  | 585/809 [11:27<04:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4669     0.1182        213        640:  73%|███████▎  | 587/809 [11:29<04:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4667     0.1184        263        640:  73%|███████▎  | 588/809 [11:30<04:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4666     0.1184        159        640:  73%|███████▎  | 590/809 [11:32<04:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4664      0.119        288        640:  73%|███████▎  | 592/809 [11:35<04:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4662      0.119        235        640:  73%|███████▎  | 594/809 [11:37<04:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4662      0.119        311        640:  74%|███████▎  | 596/809 [11:39<04:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099      0.466      0.119        294        640:  74%|███████▍  | 597/809 [11:41<04:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099      0.466      0.119        273        640:  74%|███████▍  | 599/809 [11:43<04:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4662     0.1189        204        640:  74%|███████▍  | 601/809 [11:45<04:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4661     0.1189        272        640:  74%|███████▍  | 602/809 [11:47<04:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4663     0.1187        298        640:  75%|███████▍  | 604/809 [11:49<03:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4662     0.1187        325        640:  75%|███████▍  | 606/809 [11:51<03:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4662     0.1187        248        640:  75%|███████▌  | 607/809 [11:52<03:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098      0.466     0.1188        278        640:  75%|███████▌  | 609/809 [11:55<03:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4662     0.1187        258        640:  76%|███████▌  | 611/809 [11:57<03:48,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4661     0.1186        227        640:  76%|███████▌  | 613/809 [11:59<03:45,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4662     0.1186        262        640:  76%|███████▌  | 614/809 [12:01<03:44,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4659     0.1187        266        640:  76%|███████▌  | 616/809 [12:03<03:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4658     0.1187        114        640:  76%|███████▋  | 618/809 [12:05<03:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4658     0.1189        354        640:  77%|███████▋  | 620/809 [12:08<03:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4657     0.1189        279        640:  77%|███████▋  | 621/809 [12:09<03:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4656     0.1189        488        640:  77%|███████▋  | 623/809 [12:11<03:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4657     0.1188        211        640:  77%|███████▋  | 625/809 [12:14<03:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4657     0.1187        351        640:  77%|███████▋  | 626/809 [12:15<03:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4655     0.1187        312        640:  78%|███████▊  | 628/809 [12:17<03:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4659     0.1186        281        640:  78%|███████▊  | 630/809 [12:19<03:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4658     0.1187        385        640:  78%|███████▊  | 632/809 [12:22<03:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4659     0.1186        114        640:  78%|███████▊  | 633/809 [12:23<03:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4661     0.1185        145        640:  78%|███████▊  | 635/809 [12:25<03:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4663     0.1184        154        640:  79%|███████▊  | 637/809 [12:27<03:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.097     0.4661     0.1187        407        640:  79%|███████▉  | 638/809 [12:29<03:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.098     0.4659      0.119        428        640:  79%|███████▉  | 640/809 [12:31<03:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4656      0.119        361        640:  79%|███████▉  | 642/809 [12:34<03:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4655      0.119        315        640:  79%|███████▉  | 643/809 [12:35<03:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4654      0.119        326        640:  80%|███████▉  | 645/809 [12:37<03:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1      0.465      0.119        195        640:  80%|███████▉  | 647/809 [12:40<03:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4651      0.119        162        640:  80%|████████  | 648/809 [12:41<03:19,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4651      0.119        396        640:  80%|████████  | 650/809 [12:43<03:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4652      0.119        212        640:  80%|████████  | 651/809 [12:45<03:16,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4654     0.1189        283        640:  81%|████████  | 653/809 [12:47<03:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4656      0.119        141        640:  81%|████████  | 655/809 [12:49<03:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4652     0.1191        555        640:  81%|████████  | 657/809 [12:52<03:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4653     0.1191        175        640:  81%|████████▏ | 658/809 [12:53<02:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1      0.465     0.1191        300        640:  82%|████████▏ | 660/809 [12:55<02:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4647     0.1192        326        640:  82%|████████▏ | 662/809 [12:57<02:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4647     0.1191        294        640:  82%|████████▏ | 663/809 [12:59<02:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4649     0.1191        180        640:  82%|████████▏ | 665/809 [13:01<02:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4651     0.1191        257        640:  82%|████████▏ | 667/809 [13:03<02:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1      0.465     0.1192        407        640:  83%|████████▎ | 668/809 [13:05<02:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4649     0.1193        797        640:  83%|████████▎ | 670/809 [13:07<02:49,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4648     0.1193        254        640:  83%|████████▎ | 672/809 [13:09<02:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4647     0.1192        251        640:  83%|████████▎ | 673/809 [13:11<02:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4645     0.1194        307        640:  83%|████████▎ | 675/809 [13:13<02:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4647     0.1193        280        640:  84%|████████▎ | 677/809 [13:16<02:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4647     0.1193        211        640:  84%|████████▍ | 678/809 [13:17<02:42,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4647     0.1194        315        640:  84%|████████▍ | 680/809 [13:19<02:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4647     0.1194        108        640:  84%|████████▍ | 682/809 [13:22<02:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4648     0.1194        195        640:  84%|████████▍ | 683/809 [13:23<02:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4648     0.1194         89        640:  85%|████████▍ | 685/809 [13:25<02:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4648     0.1194        377        640:  85%|████████▍ | 687/809 [13:27<02:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4645     0.1197        563        640:  85%|████████▌ | 689/809 [13:30<02:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4643     0.1197        194        640:  85%|████████▌ | 690/809 [13:31<02:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4638       0.12        803        640:  86%|████████▌ | 692/809 [13:34<02:29,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4638     0.1201        243        640:  86%|████████▌ | 693/809 [13:35<02:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4635     0.1202        292        640:  86%|████████▌ | 695/809 [13:37<02:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4638       0.12        152        640:  86%|████████▌ | 697/809 [13:39<02:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103     0.4637     0.1199        300        640:  86%|████████▋ | 699/809 [13:42<02:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4636     0.1201        263        640:  87%|████████▋ | 700/809 [13:43<02:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4637       0.12        135        640:  87%|████████▋ | 702/809 [13:45<02:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4638       0.12        307        640:  87%|████████▋ | 704/809 [13:48<02:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4636       0.12        345        640:  87%|████████▋ | 705/809 [13:49<02:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4632     0.1201        381        640:  87%|████████▋ | 707/809 [13:51<01:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4631       0.12        200        640:  88%|████████▊ | 709/809 [13:53<01:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106      0.463       0.12        247        640:  88%|████████▊ | 711/809 [13:56<01:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106     0.4628       0.12        233        640:  88%|████████▊ | 712/809 [13:57<01:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106     0.4628     0.1199        171        640:  88%|████████▊ | 714/809 [13:59<01:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106     0.4627     0.1198         94        640:  89%|████████▊ | 716/809 [14:02<01:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.107     0.4626     0.1198        231        640:  89%|████████▊ | 717/809 [14:03<01:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.107     0.4624     0.1199        180        640:  89%|████████▉ | 719/809 [14:05<01:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.108     0.4621     0.1199        247        640:  89%|████████▉ | 721/809 [14:07<01:41,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.107     0.4622     0.1197        228        640:  89%|████████▉ | 723/809 [14:10<01:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106     0.4623     0.1197        136        640:  89%|████████▉ | 724/809 [14:11<01:37,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4626     0.1195        101        640:  90%|████████▉ | 726/809 [14:13<01:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4627     0.1194        101        640:  90%|████████▉ | 728/809 [14:15<01:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4626     0.1194        270        640:  90%|█████████ | 729/809 [14:17<01:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4628     0.1193        166        640:  90%|█████████ | 731/809 [14:19<01:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4626     0.1193        436        640:  91%|█████████ | 733/809 [14:22<01:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4625     0.1193        279        640:  91%|█████████ | 734/809 [14:23<01:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4623     0.1193        158        640:  91%|█████████ | 736/809 [14:25<01:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4623     0.1192        345        640:  91%|█████████ | 738/809 [14:27<01:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4622     0.1191        254        640:  91%|█████████▏| 740/809 [14:30<01:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4622     0.1191        416        640:  92%|█████████▏| 741/809 [14:31<01:17,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106     0.4619     0.1193        523        640:  92%|█████████▏| 743/809 [14:33<01:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106      0.462     0.1192        218        640:  92%|█████████▏| 745/809 [14:36<01:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106      0.462     0.1191        224        640:  92%|█████████▏| 747/809 [14:38<01:11,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.106     0.4621     0.1191        229        640:  93%|█████████▎| 749/809 [14:40<01:08,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4621      0.119        319        640:  93%|█████████▎| 751/809 [14:42<01:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4622      0.119        170        640:  93%|█████████▎| 752/809 [14:44<01:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4623     0.1189        164        640:  93%|█████████▎| 754/809 [14:46<01:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.105     0.4623      0.119        154        640:  93%|█████████▎| 756/809 [14:48<01:00,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4624     0.1189        193        640:  94%|█████████▎| 758/809 [14:51<00:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4625     0.1188        153        640:  94%|█████████▍| 759/809 [14:52<01:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4625     0.1188        209        640:  94%|█████████▍| 761/809 [14:54<00:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4624     0.1188        145        640:  94%|█████████▍| 763/809 [14:56<00:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4623     0.1188        460        640:  94%|█████████▍| 764/809 [14:58<00:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4625     0.1187        171        640:  95%|█████████▍| 766/809 [15:00<00:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4624     0.1187        208        640:  95%|█████████▍| 768/809 [15:02<00:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103     0.4624     0.1186        290        640:  95%|█████████▌| 769/809 [15:03<00:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4627     0.1185        102        640:  95%|█████████▌| 771/809 [15:06<00:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4628     0.1184        323        640:  96%|█████████▌| 773/809 [15:08<00:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101      0.463     0.1183        395        640:  96%|█████████▌| 775/809 [15:10<00:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4627     0.1186        566        640:  96%|█████████▌| 776/809 [15:12<00:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4626     0.1187        622        640:  96%|█████████▌| 778/809 [15:14<00:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103     0.4622      0.119        458        640:  96%|█████████▋| 780/809 [15:16<00:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4621      0.119        261        640:  97%|█████████▋| 781/809 [15:18<00:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104     0.4621     0.1189        394        640:  97%|█████████▋| 783/809 [15:20<00:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.104      0.462     0.1191        297        640:  97%|█████████▋| 785/809 [15:22<00:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103     0.4621      0.119        131        640:  97%|█████████▋| 786/809 [15:23<00:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.103     0.4623     0.1189        152        640:  97%|█████████▋| 788/809 [15:26<00:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4625     0.1188        157        640:  98%|█████████▊| 790/809 [15:28<00:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.102     0.4625     0.1188        333        640:  98%|█████████▊| 792/809 [15:31<00:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4626     0.1187        181        640:  98%|█████████▊| 793/809 [15:32<00:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4627     0.1187        186        640:  98%|█████████▊| 795/809 [15:34<00:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.101     0.4626     0.1187        435        640:  99%|█████████▊| 797/809 [15:36<00:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4628     0.1186         49        640:  99%|█████████▊| 798/809 [15:38<00:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4628     0.1187        433        640:  99%|█████████▉| 800/809 [15:40<00:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4628     0.1186        126        640:  99%|█████████▉| 802/809 [15:42<00:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099     0.4629     0.1186        255        640:  99%|█████████▉| 804/809 [15:45<00:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G      1.099      0.463     0.1185        195        640: 100%|█████████▉| 805/809 [15:46<00:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     71/300      13.8G        1.1     0.4628     0.1185        407        640: 100%|█████████▉| 807/809 [15:48<00:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   1%|▏         | 1/69 [00:00<00:18,  3.75it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  13%|█▎        | 9/69 [00:02<00:14,  4.17it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  25%|██▍       | 17/69 [00:04<00:13,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  36%|███▌      | 25/69 [00:06<00:10,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  48%|████▊     | 33/69 [00:08<00:08,  4.04it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  59%|█████▉    | 41/69 [00:10<00:06,  4.08it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  71%|███████   | 49/69 [00:12<00:05,  3.73it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  83%|████████▎ | 57/69 [00:14<00:02,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:00,  4.07it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G     0.7547     0.5633    0.06234        168        640:   0%|          | 1/809 [00:01<15:07,  1.12s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.06     0.4588    0.08864        338        640:   0%|          | 3/809 [00:03<15:23,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.137      0.443     0.1057        316        640:   1%|          | 5/809 [00:05<15:13,  1.14s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.181     0.4304     0.1162        603        640:   1%|          | 7/809 [00:07<15:18,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.129     0.4534     0.1132        182        640:   1%|          | 8/809 [00:09<15:26,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.117      0.459     0.1115        200        640:   1%|          | 10/809 [00:11<15:34,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.083     0.4654     0.1061        234        640:   1%|▏         | 12/809 [00:13<15:43,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.084     0.4669     0.1045        214        640:   2%|▏         | 13/809 [00:15<15:46,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.137     0.4505     0.1263        316        640:   2%|▏         | 15/809 [00:17<16:14,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.12     0.4554     0.1221        206        640:   2%|▏         | 17/809 [00:20<15:41,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.104      0.461     0.1189         87        640:   2%|▏         | 18/809 [00:21<17:28,  1.33s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.093     0.4627     0.1168        377        640:   2%|▏         | 20/809 [00:23<16:19,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.088     0.4629     0.1155        208        640:   3%|▎         | 21/809 [00:25<16:01,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.076     0.4666     0.1131        233        640:   3%|▎         | 23/809 [00:27<15:49,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.061     0.4712     0.1119        119        640:   3%|▎         | 25/809 [00:29<15:42,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.051     0.4721     0.1106        227        640:   3%|▎         | 26/809 [00:31<15:35,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.036     0.4757     0.1088        179        640:   3%|▎         | 28/809 [00:33<15:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.038     0.4743     0.1072        323        640:   4%|▎         | 30/809 [00:35<15:30,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.036     0.4752     0.1066        191        640:   4%|▍         | 31/809 [00:37<15:24,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.057     0.4689     0.1097        305        640:   4%|▍         | 33/809 [00:39<15:30,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.057     0.4695     0.1089        421        640:   4%|▍         | 35/809 [00:41<15:40,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.055     0.4696     0.1083        255        640:   4%|▍         | 36/809 [00:43<15:28,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.064     0.4664     0.1089        352        640:   5%|▍         | 38/809 [00:45<15:20,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.077     0.4617     0.1103        349        640:   5%|▍         | 40/809 [00:47<15:17,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.078     0.4613     0.1105        221        640:   5%|▌         | 42/809 [00:50<15:06,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.086     0.4595     0.1112        275        640:   5%|▌         | 43/809 [00:51<14:50,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.088     0.4575     0.1121        272        640:   6%|▌         | 45/809 [00:53<14:38,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.079      0.459      0.111        243        640:   6%|▌         | 47/809 [00:55<14:46,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.078      0.458     0.1112        385        640:   6%|▌         | 49/809 [00:58<14:38,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.083     0.4561     0.1117        474        640:   6%|▌         | 50/809 [00:59<14:42,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.081     0.4567     0.1109        311        640:   6%|▋         | 52/809 [01:01<14:29,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.084     0.4565      0.111        443        640:   7%|▋         | 54/809 [01:04<14:45,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.084     0.4562     0.1106        267        640:   7%|▋         | 55/809 [01:05<14:39,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.09     0.4546       0.11        398        640:   7%|▋         | 57/809 [01:07<14:35,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.096     0.4525       0.11        458        640:   7%|▋         | 59/809 [01:09<14:28,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.099     0.4526     0.1098        226        640:   8%|▊         | 61/809 [01:12<14:26,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.105     0.4506       0.11        267        640:   8%|▊         | 62/809 [01:13<14:38,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.103     0.4519     0.1095        202        640:   8%|▊         | 64/809 [01:15<14:57,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.118      0.448     0.1118        411        640:   8%|▊         | 66/809 [01:18<14:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.114     0.4491     0.1116        165        640:   8%|▊         | 67/809 [01:19<14:38,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.125     0.4466     0.1144        310        640:   9%|▊         | 69/809 [01:21<14:41,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.121     0.4472     0.1142         56        640:   9%|▉         | 71/809 [01:24<15:05,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.12      0.448     0.1139        147        640:   9%|▉         | 72/809 [01:25<14:51,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.119     0.4484     0.1134        175        640:   9%|▉         | 74/809 [01:27<15:14,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.117      0.449     0.1129        161        640:   9%|▉         | 76/809 [01:30<14:37,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.115     0.4495     0.1124        198        640:  10%|▉         | 77/809 [01:31<14:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.119     0.4488     0.1121        524        640:  10%|▉         | 79/809 [01:33<14:27,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.121     0.4487     0.1124        312        640:  10%|█         | 81/809 [01:36<14:10,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.121     0.4488     0.1123        363        640:  10%|█         | 82/809 [01:37<14:02,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.122      0.449     0.1127        141        640:  10%|█         | 84/809 [01:39<14:04,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.118     0.4501     0.1128        228        640:  11%|█         | 86/809 [01:41<14:06,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.115     0.4504      0.112        187        640:  11%|█         | 88/809 [01:44<14:13,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.119     0.4491     0.1127        332        640:  11%|█         | 90/809 [01:46<13:53,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.124     0.4485     0.1134        177        640:  11%|█▏        | 92/809 [01:48<13:50,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.126     0.4481     0.1132        238        640:  11%|█▏        | 93/809 [01:50<13:48,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.125     0.4486     0.1129        306        640:  12%|█▏        | 95/809 [01:52<14:06,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.13     0.4476     0.1135        183        640:  12%|█▏        | 97/809 [01:54<13:52,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4471     0.1132        228        640:  12%|█▏        | 98/809 [01:56<13:54,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.133      0.447     0.1134        293        640:  12%|█▏        | 100/809 [01:58<13:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.134     0.4467     0.1131        234        640:  13%|█▎        | 102/809 [02:00<13:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.135     0.4467     0.1134        244        640:  13%|█▎        | 104/809 [02:02<13:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.137     0.4463     0.1137        292        640:  13%|█▎        | 105/809 [02:04<13:28,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.139     0.4454      0.114        129        640:  13%|█▎        | 107/809 [02:06<13:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.136     0.4464     0.1136        226        640:  13%|█▎        | 109/809 [02:08<13:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.136      0.446     0.1135        271        640:  14%|█▎        | 110/809 [02:09<13:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4475      0.113        113        640:  14%|█▍        | 112/809 [02:12<13:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.134     0.4464     0.1135        247        640:  14%|█▍        | 114/809 [02:14<13:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.132     0.4476     0.1132        189        640:  14%|█▍        | 115/809 [02:15<13:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.137     0.4462     0.1152        214        640:  14%|█▍        | 117/809 [02:18<13:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.13     0.4482     0.1145         93        640:  15%|█▍        | 119/809 [02:20<13:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.129     0.4486     0.1147        151        640:  15%|█▍        | 121/809 [02:23<13:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.126     0.4492     0.1143        116        640:  15%|█▌        | 122/809 [02:24<13:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.134     0.4472     0.1185        516        640:  15%|█▌        | 124/809 [02:26<13:54,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.137     0.4465     0.1201        406        640:  15%|█▌        | 125/809 [02:27<13:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.132     0.4475     0.1196        131        640:  16%|█▌        | 127/809 [02:30<13:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4472     0.1191        192        640:  16%|█▌        | 129/809 [02:32<13:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.135      0.446      0.121        449        640:  16%|█▌        | 131/809 [02:35<13:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.135     0.4457     0.1208        222        640:  16%|█▋        | 132/809 [02:36<13:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.138     0.4448     0.1228        265        640:  17%|█▋        | 134/809 [02:38<13:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4421     0.1236        523        640:  17%|█▋        | 136/809 [02:40<12:56,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.148     0.4412     0.1236        177        640:  17%|█▋        | 137/809 [02:41<12:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4396     0.1236        190        640:  17%|█▋        | 139/809 [02:44<13:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4375     0.1246        450        640:  17%|█▋        | 141/809 [02:46<13:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4366     0.1247        136        640:  18%|█▊        | 143/809 [02:49<13:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158      0.437     0.1246        177        640:  18%|█▊        | 144/809 [02:50<13:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4369     0.1244        282        640:  18%|█▊        | 146/809 [02:52<12:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.154     0.4375     0.1242        196        640:  18%|█▊        | 148/809 [02:54<12:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.154     0.4376     0.1241        369        640:  18%|█▊        | 149/809 [02:56<12:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.154     0.4372     0.1238        254        640:  19%|█▊        | 151/809 [02:58<12:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.156     0.4367     0.1239        308        640:  19%|█▉        | 153/809 [03:00<12:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4356     0.1243        294        640:  19%|█▉        | 155/809 [03:03<12:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4353     0.1241        193        640:  19%|█▉        | 156/809 [03:04<12:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4349     0.1238        194        640:  20%|█▉        | 158/809 [03:06<12:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4353     0.1235         69        640:  20%|█▉        | 160/809 [03:09<13:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4357     0.1233        117        640:  20%|█▉        | 161/809 [03:10<12:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4354     0.1266        456        640:  20%|██        | 163/809 [03:12<12:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4351     0.1278        425        640:  20%|██        | 165/809 [03:14<12:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.164     0.4344     0.1294        730        640:  21%|██        | 166/809 [03:16<12:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.166     0.4339     0.1292        250        640:  21%|██        | 168/809 [03:18<12:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.164     0.4339     0.1286        181        640:  21%|██        | 170/809 [03:20<12:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4342     0.1282        145        640:  21%|██▏       | 172/809 [03:23<12:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163      0.434     0.1282        230        640:  21%|██▏       | 173/809 [03:24<12:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.164     0.4335     0.1279        170        640:  22%|██▏       | 175/809 [03:26<12:10,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.164     0.4335     0.1277        264        640:  22%|██▏       | 177/809 [03:28<12:06,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163      0.434     0.1276        197        640:  22%|██▏       | 179/809 [03:31<12:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163     0.4339     0.1274        228        640:  22%|██▏       | 180/809 [03:32<12:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4342     0.1274        384        640:  22%|██▏       | 182/809 [03:34<12:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4352     0.1268        155        640:  23%|██▎       | 184/809 [03:36<12:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4351     0.1268        213        640:  23%|██▎       | 185/809 [03:38<11:57,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.156     0.4357     0.1265        238        640:  23%|██▎       | 187/809 [03:40<12:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4356     0.1263        149        640:  23%|██▎       | 189/809 [03:42<11:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152      0.436     0.1258        199        640:  24%|██▎       | 191/809 [03:45<12:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4352     0.1266        355        640:  24%|██▎       | 192/809 [03:46<12:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.154     0.4353     0.1263        179        640:  24%|██▍       | 194/809 [03:48<12:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.156     0.4348     0.1264        253        640:  24%|██▍       | 196/809 [03:51<11:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.154     0.4347     0.1263        180        640:  24%|██▍       | 197/809 [03:52<11:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.153     0.4351      0.126        176        640:  25%|██▍       | 199/809 [03:54<11:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4357     0.1257        201        640:  25%|██▍       | 201/809 [03:56<11:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.153     0.4353      0.126        355        640:  25%|██▍       | 202/809 [03:58<12:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4356     0.1258        319        640:  25%|██▌       | 204/809 [04:00<12:23,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.15     0.4362     0.1256        180        640:  25%|██▌       | 206/809 [04:03<12:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4357     0.1256        241        640:  26%|██▌       | 207/809 [04:04<12:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4347      0.126        561        640:  26%|██▌       | 209/809 [04:06<12:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.156     0.4349      0.126        294        640:  26%|██▌       | 211/809 [04:09<11:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4355     0.1258        138        640:  26%|██▌       | 212/809 [04:10<11:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.153     0.4361     0.1255        106        640:  26%|██▋       | 214/809 [04:12<11:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4368     0.1252        212        640:  27%|██▋       | 216/809 [04:14<11:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.15     0.4373     0.1248        157        640:  27%|██▋       | 218/809 [04:17<11:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149     0.4376     0.1246        169        640:  27%|██▋       | 219/809 [04:18<11:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149     0.4379     0.1244        183        640:  27%|██▋       | 221/809 [04:20<11:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149     0.4379     0.1242        285        640:  28%|██▊       | 223/809 [04:22<11:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147     0.4389     0.1239         89        640:  28%|██▊       | 225/809 [04:25<11:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4392     0.1237        274        640:  28%|██▊       | 226/809 [04:26<11:12,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4394     0.1237        454        640:  28%|██▊       | 228/809 [04:28<11:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146       0.44     0.1235        233        640:  28%|██▊       | 230/809 [04:31<11:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.145     0.4402     0.1235         40        640:  29%|██▊       | 231/809 [04:32<11:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4397     0.1234        266        640:  29%|██▉       | 233/809 [04:34<11:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149      0.439     0.1234        253        640:  29%|██▉       | 235/809 [04:37<11:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4386     0.1234        213        640:  29%|██▉       | 237/809 [04:39<10:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.15     0.4388     0.1233        156        640:  29%|██▉       | 238/809 [04:40<11:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.148     0.4396     0.1231        141        640:  30%|██▉       | 240/809 [04:42<11:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147     0.4399     0.1231        194        640:  30%|██▉       | 242/809 [04:45<11:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147     0.4398     0.1232        337        640:  30%|███       | 243/809 [04:46<11:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.148     0.4397     0.1233        363        640:  30%|███       | 245/809 [04:48<11:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149     0.4396     0.1237        314        640:  31%|███       | 247/809 [04:51<11:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149     0.4404     0.1235        265        640:  31%|███       | 249/809 [04:53<10:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.148     0.4406     0.1234        211        640:  31%|███       | 250/809 [04:54<11:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151       0.44      0.124        453        640:  31%|███       | 252/809 [04:57<11:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151       0.44     0.1238        293        640:  31%|███▏      | 254/809 [04:59<10:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149     0.4404     0.1237        151        640:  32%|███▏      | 255/809 [05:00<10:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.148      0.441     0.1235        348        640:  32%|███▏      | 257/809 [05:02<10:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147      0.441     0.1233        412        640:  32%|███▏      | 259/809 [05:05<10:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.148     0.4409     0.1233        314        640:  32%|███▏      | 261/809 [05:07<10:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147      0.441     0.1232        293        640:  32%|███▏      | 262/809 [05:09<11:22,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4414     0.1231        234        640:  33%|███▎      | 264/809 [05:11<10:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147     0.4412     0.1233        412        640:  33%|███▎      | 266/809 [05:13<10:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.145     0.4417     0.1229        272        640:  33%|███▎      | 268/809 [05:15<10:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4414     0.1231        412        640:  33%|███▎      | 269/809 [05:17<10:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4415     0.1234        294        640:  33%|███▎      | 271/809 [05:19<10:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.145     0.4417     0.1232        166        640:  34%|███▎      | 273/809 [05:21<10:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4416     0.1231        313        640:  34%|███▍      | 275/809 [05:24<10:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4415     0.1231        277        640:  34%|███▍      | 276/809 [05:25<10:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.145     0.4416      0.123        193        640:  34%|███▍      | 278/809 [05:27<10:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.145     0.4416     0.1229        239        640:  35%|███▍      | 280/809 [05:30<10:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4413     0.1229        322        640:  35%|███▍      | 281/809 [05:31<10:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.148     0.4407     0.1237        448        640:  35%|███▍      | 283/809 [05:33<10:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4409     0.1237        277        640:  35%|███▌      | 285/809 [05:35<10:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.145     0.4411     0.1237        337        640:  35%|███▌      | 286/809 [05:37<10:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.145     0.4414     0.1236        371        640:  36%|███▌      | 288/809 [05:39<10:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.145     0.4418     0.1234        320        640:  36%|███▌      | 290/809 [05:41<10:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4413     0.1237        197        640:  36%|███▌      | 292/809 [05:44<10:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147      0.441     0.1238        506        640:  36%|███▌      | 293/809 [05:45<10:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147     0.4412     0.1239        154        640:  36%|███▋      | 295/809 [05:47<10:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147     0.4414      0.124        201        640:  37%|███▋      | 297/809 [05:50<10:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.148      0.441     0.1242        259        640:  37%|███▋      | 298/809 [05:51<09:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149      0.441     0.1245        141        640:  37%|███▋      | 300/809 [05:53<10:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.148     0.4413     0.1244        215        640:  37%|███▋      | 302/809 [05:56<10:20,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149     0.4411     0.1245        228        640:  37%|███▋      | 303/809 [05:57<10:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149     0.4409     0.1244        264        640:  38%|███▊      | 305/809 [05:59<10:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.15     0.4408     0.1245        147        640:  38%|███▊      | 307/809 [06:01<09:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4406     0.1246        273        640:  38%|███▊      | 308/809 [06:03<10:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4406     0.1246        128        640:  38%|███▊      | 310/809 [06:05<10:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4402     0.1247        328        640:  39%|███▊      | 312/809 [06:08<10:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4405     0.1247        128        640:  39%|███▊      | 313/809 [06:09<09:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4404     0.1247        371        640:  39%|███▉      | 315/809 [06:11<09:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4405     0.1246        150        640:  39%|███▉      | 317/809 [06:13<09:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4402     0.1246        312        640:  39%|███▉      | 318/809 [06:15<09:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4399     0.1247        178        640:  40%|███▉      | 320/809 [06:17<09:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.154     0.4396     0.1258        226        640:  40%|███▉      | 322/809 [06:19<09:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.156     0.4389     0.1266        503        640:  40%|███▉      | 323/809 [06:21<09:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4383     0.1271        408        640:  40%|████      | 325/809 [06:23<09:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4384     0.1271        200        640:  40%|████      | 327/809 [06:25<09:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4382     0.1272        416        640:  41%|████      | 329/809 [06:28<09:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4379     0.1271        178        640:  41%|████      | 330/809 [06:29<09:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4379     0.1273        215        640:  41%|████      | 332/809 [06:31<09:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4374     0.1276        228        640:  41%|████▏     | 334/809 [06:33<09:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4376     0.1276        345        640:  42%|████▏     | 336/809 [06:36<09:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4375     0.1276        138        640:  42%|████▏     | 337/809 [06:37<09:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4376     0.1275        258        640:  42%|████▏     | 339/809 [06:39<09:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158      0.438     0.1273        260        640:  42%|████▏     | 341/809 [06:42<09:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4376     0.1277        341        640:  42%|████▏     | 342/809 [06:43<09:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4378     0.1276        345        640:  43%|████▎     | 344/809 [06:45<09:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4377     0.1283        552        640:  43%|████▎     | 346/809 [06:48<09:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4379     0.1282        185        640:  43%|████▎     | 348/809 [06:50<08:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4381     0.1282        230        640:  43%|████▎     | 349/809 [06:51<08:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4382      0.128        205        640:  43%|████▎     | 351/809 [06:53<08:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4381     0.1281        267        640:  44%|████▎     | 353/809 [06:56<08:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4382      0.128        331        640:  44%|████▍     | 354/809 [06:57<08:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4385      0.128        333        640:  44%|████▍     | 356/809 [06:59<08:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.156     0.4388     0.1279        465        640:  44%|████▍     | 358/809 [07:02<08:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.156      0.439     0.1278        281        640:  44%|████▍     | 360/809 [07:04<08:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4391     0.1277        268        640:  45%|████▍     | 361/809 [07:05<08:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.153       0.44     0.1275        115        640:  45%|████▍     | 363/809 [07:07<08:34,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4403     0.1274        186        640:  45%|████▌     | 365/809 [07:10<08:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4404     0.1274        309        640:  45%|████▌     | 366/809 [07:11<08:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4399     0.1275        206        640:  45%|████▌     | 368/809 [07:13<08:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4399     0.1273        173        640:  46%|████▌     | 370/809 [07:16<08:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4398     0.1273        197        640:  46%|████▌     | 371/809 [07:17<08:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4398     0.1272        297        640:  46%|████▌     | 373/809 [07:19<08:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4401     0.1271        244        640:  46%|████▋     | 375/809 [07:22<08:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4402     0.1272        177        640:  47%|████▋     | 377/809 [07:24<08:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4399     0.1271        395        640:  47%|████▋     | 378/809 [07:25<08:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151       0.44      0.127        327        640:  47%|████▋     | 380/809 [07:27<08:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.15     0.4402     0.1269        285        640:  47%|████▋     | 382/809 [07:30<08:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.15     0.4403     0.1268        155        640:  47%|████▋     | 383/809 [07:31<08:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.15     0.4401     0.1267        181        640:  48%|████▊     | 385/809 [07:33<08:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151       0.44      0.127        159        640:  48%|████▊     | 387/809 [07:36<08:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4396     0.1271        488        640:  48%|████▊     | 388/809 [07:37<08:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.153     0.4393     0.1274        641        640:  48%|████▊     | 390/809 [07:40<08:34,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.153     0.4392     0.1272        289        640:  48%|████▊     | 392/809 [07:42<08:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4387     0.1279        365        640:  49%|████▊     | 393/809 [07:43<08:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4389     0.1284         59        640:  49%|████▉     | 395/809 [07:45<08:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155      0.439     0.1285        401        640:  49%|████▉     | 397/809 [07:48<08:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4392     0.1284        189        640:  49%|████▉     | 398/809 [07:49<08:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4386     0.1287        236        640:  49%|████▉     | 400/809 [07:51<08:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4382     0.1287        317        640:  50%|████▉     | 402/809 [07:54<07:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4379     0.1288        292        640:  50%|████▉     | 404/809 [07:56<07:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159      0.438     0.1287        259        640:  50%|█████     | 405/809 [07:57<07:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4381     0.1286        299        640:  50%|█████     | 407/809 [08:00<07:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4378     0.1287        319        640:  51%|█████     | 409/809 [08:02<07:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4377     0.1288        254        640:  51%|█████     | 410/809 [08:03<07:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4376     0.1288        335        640:  51%|█████     | 413/809 [08:07<07:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4373     0.1291        514        640:  51%|█████     | 414/809 [08:08<07:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4369     0.1293        582        640:  51%|█████▏    | 416/809 [08:10<07:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4371     0.1291        335        640:  52%|█████▏    | 418/809 [08:12<07:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4371     0.1291        149        640:  52%|█████▏    | 419/809 [08:14<07:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4371     0.1293        338        640:  52%|█████▏    | 421/809 [08:16<07:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4369     0.1292        276        640:  52%|█████▏    | 423/809 [08:18<07:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4369     0.1292        157        640:  52%|█████▏    | 424/809 [08:20<07:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4368     0.1292        270        640:  53%|█████▎    | 426/809 [08:22<07:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4368     0.1291        209        640:  53%|█████▎    | 428/809 [08:24<07:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162      0.437      0.129        221        640:  53%|█████▎    | 429/809 [08:26<07:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4371     0.1289        105        640:  53%|█████▎    | 431/809 [08:28<07:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4374     0.1287        109        640:  54%|█████▎    | 433/809 [08:30<07:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4373     0.1287        223        640:  54%|█████▍    | 435/809 [08:33<07:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4372     0.1287        331        640:  54%|█████▍    | 436/809 [08:34<07:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4371     0.1285        261        640:  54%|█████▍    | 438/809 [08:36<07:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162      0.437     0.1285        238        640:  54%|█████▍    | 440/809 [08:39<07:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4369     0.1284        173        640:  55%|█████▍    | 441/809 [08:40<07:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.164     0.4364     0.1289        455        640:  55%|█████▍    | 443/809 [08:42<07:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.164     0.4364     0.1289        508        640:  55%|█████▌    | 445/809 [08:44<07:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.164     0.4363     0.1289        213        640:  55%|█████▌    | 446/809 [08:46<07:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163     0.4363     0.1288        235        640:  55%|█████▌    | 448/809 [08:48<07:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163     0.4361     0.1287        290        640:  56%|█████▌    | 450/809 [08:50<06:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.164     0.4359     0.1287        259        640:  56%|█████▌    | 452/809 [08:53<06:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163      0.436     0.1287        341        640:  56%|█████▌    | 453/809 [08:54<06:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163      0.436     0.1286        367        640:  56%|█████▌    | 455/809 [08:56<07:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4361     0.1284        320        640:  56%|█████▋    | 457/809 [08:58<06:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4362     0.1283        485        640:  57%|█████▋    | 458/809 [09:00<06:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4364     0.1282        107        640:  57%|█████▋    | 460/809 [09:02<06:42,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4366      0.128        326        640:  57%|█████▋    | 462/809 [09:04<06:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4366     0.1283        316        640:  57%|█████▋    | 464/809 [09:07<06:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4368     0.1282        185        640:  57%|█████▋    | 465/809 [09:08<06:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4371     0.1281        244        640:  58%|█████▊    | 467/809 [09:10<06:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4371      0.128        343        640:  58%|█████▊    | 469/809 [09:13<06:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159      0.437     0.1279        258        640:  58%|█████▊    | 470/809 [09:14<06:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4373     0.1277        208        640:  58%|█████▊    | 472/809 [09:16<06:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4373     0.1276        292        640:  59%|█████▊    | 474/809 [09:18<06:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4375     0.1274        239        640:  59%|█████▉    | 476/809 [09:21<06:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4372     0.1275        516        640:  59%|█████▉    | 477/809 [09:22<06:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4372     0.1275        531        640:  59%|█████▉    | 479/809 [09:24<06:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4376     0.1272        131        640:  59%|█████▉    | 481/809 [09:27<06:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4379      0.127        425        640:  60%|█████▉    | 483/809 [09:29<06:14,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4379     0.1271        306        640:  60%|█████▉    | 484/809 [09:30<06:12,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4377     0.1274        319        640:  60%|██████    | 486/809 [09:32<06:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4376     0.1273        288        640:  60%|██████    | 488/809 [09:35<06:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4373     0.1275        457        640:  60%|██████    | 489/809 [09:36<06:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4374     0.1274        269        640:  61%|██████    | 491/809 [09:38<06:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4373     0.1274        184        640:  61%|██████    | 493/809 [09:40<06:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4373     0.1273        306        640:  61%|██████    | 495/809 [09:43<06:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4375     0.1273        139        640:  61%|██████▏   | 496/809 [09:44<06:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158      0.438     0.1271         97        640:  62%|██████▏   | 498/809 [09:46<06:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4379      0.127        155        640:  62%|██████▏   | 500/809 [09:49<06:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4379     0.1269        308        640:  62%|██████▏   | 501/809 [09:50<06:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.156     0.4382     0.1268         98        640:  62%|██████▏   | 503/809 [09:52<06:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4381     0.1268        309        640:  62%|██████▏   | 504/809 [09:54<06:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.156     0.4381     0.1266        274        640:  63%|██████▎   | 506/809 [09:56<06:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4385     0.1266        131        640:  63%|██████▎   | 508/809 [09:58<05:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.154     0.4387     0.1265        216        640:  63%|██████▎   | 510/809 [10:01<05:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.153     0.4389     0.1264         79        640:  63%|██████▎   | 511/809 [10:02<05:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4392     0.1263        187        640:  63%|██████▎   | 513/809 [10:04<05:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4392     0.1262        306        640:  64%|██████▎   | 515/809 [10:07<05:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4392     0.1261        301        640:  64%|██████▍   | 516/809 [10:08<05:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4393      0.126        186        640:  64%|██████▍   | 518/809 [10:10<05:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152     0.4393     0.1261        380        640:  64%|██████▍   | 520/809 [10:13<05:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.153      0.439      0.126        225        640:  65%|██████▍   | 522/809 [10:15<05:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.154     0.4387     0.1262        329        640:  65%|██████▍   | 523/809 [10:16<05:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.154     0.4387      0.126        292        640:  65%|██████▍   | 525/809 [10:18<05:24,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4383     0.1261        374        640:  65%|██████▌   | 527/809 [10:21<05:23,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.156      0.438     0.1261        235        640:  65%|██████▌   | 529/809 [10:23<05:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4379     0.1261        369        640:  66%|██████▌   | 530/809 [10:24<05:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4377     0.1261        176        640:  66%|██████▌   | 532/809 [10:26<05:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4375     0.1262        329        640:  66%|██████▌   | 534/809 [10:29<05:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4375     0.1261        221        640:  66%|██████▌   | 535/809 [10:30<05:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4376     0.1261        312        640:  66%|██████▋   | 537/809 [10:32<05:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4374     0.1261        460        640:  67%|██████▋   | 539/809 [10:35<05:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4373      0.126        320        640:  67%|██████▋   | 541/809 [10:37<05:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4374      0.126        313        640:  67%|██████▋   | 542/809 [10:38<05:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4375     0.1258        238        640:  67%|██████▋   | 544/809 [10:41<05:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4374     0.1259        327        640:  67%|██████▋   | 546/809 [10:43<05:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4372     0.1258        388        640:  68%|██████▊   | 547/809 [10:44<05:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4373     0.1257        318        640:  68%|██████▊   | 549/809 [10:46<05:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16      0.437     0.1256        272        640:  68%|██████▊   | 551/809 [10:49<05:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159      0.437     0.1256        366        640:  68%|██████▊   | 553/809 [10:51<04:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4372     0.1255        263        640:  68%|██████▊   | 554/809 [10:52<04:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4373     0.1255        426        640:  69%|██████▊   | 556/809 [10:55<04:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4375     0.1254        193        640:  69%|██████▉   | 558/809 [10:57<04:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4376     0.1253        178        640:  69%|██████▉   | 559/809 [10:58<04:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4374     0.1253        249        640:  69%|██████▉   | 561/809 [11:00<04:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4369     0.1255        295        640:  70%|██████▉   | 563/809 [11:03<04:42,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4367     0.1255        176        640:  70%|██████▉   | 565/809 [11:05<04:38,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4366     0.1254        180        640:  70%|███████   | 567/809 [11:07<04:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4369     0.1253        330        640:  70%|███████   | 569/809 [11:10<04:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16      0.437     0.1253        301        640:  70%|███████   | 570/809 [11:11<04:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4369     0.1252        346        640:  71%|███████   | 572/809 [11:13<04:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4368     0.1252        455        640:  71%|███████   | 574/809 [11:15<04:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4367     0.1252        262        640:  71%|███████   | 575/809 [11:17<04:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4368     0.1252         93        640:  71%|███████▏  | 577/809 [11:19<04:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4365     0.1252        250        640:  72%|███████▏  | 579/809 [11:21<04:25,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163     0.4361     0.1252        206        640:  72%|███████▏  | 581/809 [11:24<04:22,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4361     0.1252        316        640:  72%|███████▏  | 582/809 [11:25<04:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4363     0.1251        321        640:  72%|███████▏  | 584/809 [11:27<04:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4364      0.125        443        640:  72%|███████▏  | 586/809 [11:29<04:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4365     0.1249        302        640:  73%|███████▎  | 587/809 [11:31<04:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4368     0.1248        163        640:  73%|███████▎  | 589/809 [11:33<04:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4366     0.1248        443        640:  73%|███████▎  | 591/809 [11:35<04:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4365      0.125        173        640:  73%|███████▎  | 593/809 [11:38<04:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4366     0.1249        200        640:  73%|███████▎  | 594/809 [11:39<04:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163     0.4363     0.1251        571        640:  74%|███████▎  | 596/809 [11:41<04:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163     0.4363     0.1251        343        640:  74%|███████▍  | 598/809 [11:43<04:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163     0.4364     0.1251        147        640:  74%|███████▍  | 599/809 [11:45<04:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163     0.4362     0.1251        720        640:  74%|███████▍  | 601/809 [11:47<04:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163     0.4363     0.1251        245        640:  75%|███████▍  | 603/809 [11:49<04:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4363     0.1251        437        640:  75%|███████▍  | 604/809 [11:51<04:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163     0.4363     0.1251        557        640:  75%|███████▍  | 606/809 [11:53<04:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4362      0.125        266        640:  75%|███████▌  | 608/809 [11:55<03:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.163     0.4362     0.1251        630        640:  75%|███████▌  | 609/809 [11:57<03:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.162     0.4363      0.125        321        640:  76%|███████▌  | 611/809 [11:59<03:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4368      0.125        153        640:  76%|███████▌  | 613/809 [12:01<03:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4369     0.1251        343        640:  76%|███████▌  | 615/809 [12:04<03:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.161     0.4368     0.1252        318        640:  76%|███████▌  | 616/809 [12:05<03:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.16     0.4373      0.125        174        640:  76%|███████▋  | 618/809 [12:07<03:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4375     0.1249        255        640:  77%|███████▋  | 620/809 [12:09<03:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4376     0.1251        283        640:  77%|███████▋  | 622/809 [12:12<03:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.159     0.4375     0.1251        376        640:  77%|███████▋  | 623/809 [12:13<03:34,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4376      0.125        305        640:  77%|███████▋  | 625/809 [12:15<03:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4377     0.1249        231        640:  78%|███████▊  | 627/809 [12:17<03:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.158     0.4379      0.125        412        640:  78%|███████▊  | 629/809 [12:20<03:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157      0.438      0.125        201        640:  78%|███████▊  | 630/809 [12:21<03:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.157     0.4381     0.1249        183        640:  78%|███████▊  | 632/809 [12:23<03:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4383     0.1248        177        640:  78%|███████▊  | 634/809 [12:26<03:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.155     0.4385     0.1248        149        640:  78%|███████▊  | 635/809 [12:27<03:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.154     0.4386     0.1247        165        640:  79%|███████▊  | 637/809 [12:29<03:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.153     0.4389     0.1246        194        640:  79%|███████▉  | 639/809 [12:32<03:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.152      0.439     0.1245        123        640:  79%|███████▉  | 640/809 [12:33<03:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.151     0.4392     0.1245        237        640:  79%|███████▉  | 642/809 [12:35<03:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.15     0.4396     0.1244        133        640:  80%|███████▉  | 644/809 [12:37<03:10,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149     0.4399     0.1243        129        640:  80%|███████▉  | 646/809 [12:40<03:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149       0.44     0.1242        447        640:  80%|███████▉  | 647/809 [12:41<03:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149       0.44     0.1242        475        640:  80%|████████  | 649/809 [12:43<03:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.148     0.4401     0.1244        550        640:  80%|████████  | 651/809 [12:46<03:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149     0.4399     0.1246        547        640:  81%|████████  | 652/809 [12:47<03:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.149       0.44     0.1247        386        640:  81%|████████  | 654/809 [12:49<03:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.148     0.4401     0.1246        244        640:  81%|████████  | 656/809 [12:52<03:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147     0.4403     0.1245        116        640:  81%|████████  | 657/809 [12:53<03:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147     0.4404     0.1244        485        640:  81%|████████▏ | 659/809 [12:55<02:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4404     0.1244        316        640:  82%|████████▏ | 661/809 [12:58<02:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.147     0.4404     0.1245        410        640:  82%|████████▏ | 662/809 [12:59<02:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4405     0.1244        308        640:  82%|████████▏ | 664/809 [13:01<02:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4406     0.1243        327        640:  82%|████████▏ | 666/809 [13:03<02:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.145     0.4408     0.1243        334        640:  83%|████████▎ | 668/809 [13:06<02:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.146     0.4406     0.1243        273        640:  83%|████████▎ | 669/809 [13:07<02:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.145      0.441     0.1241        137        640:  83%|████████▎ | 671/809 [13:09<02:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.144     0.4413      0.124        339        640:  83%|████████▎ | 673/809 [13:12<02:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.143     0.4415     0.1239        347        640:  83%|████████▎ | 675/809 [13:14<02:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.143     0.4416     0.1238        231        640:  84%|████████▎ | 676/809 [13:15<02:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.143     0.4417     0.1237        292        640:  84%|████████▍ | 678/809 [13:17<02:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.142     0.4419     0.1236        214        640:  84%|████████▍ | 680/809 [13:20<02:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.143     0.4418     0.1236        381        640:  84%|████████▍ | 681/809 [13:21<02:27,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.143     0.4418     0.1236        318        640:  84%|████████▍ | 683/809 [13:23<02:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.143     0.4417     0.1238        153        640:  85%|████████▍ | 685/809 [13:26<02:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.142      0.442     0.1236        339        640:  85%|████████▍ | 687/809 [13:28<02:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.142      0.442     0.1236        132        640:  85%|████████▌ | 688/809 [13:29<02:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.142     0.4421     0.1235        307        640:  85%|████████▌ | 690/809 [13:31<02:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141     0.4423     0.1233        183        640:  86%|████████▌ | 692/809 [13:34<02:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141     0.4423     0.1234        631        640:  86%|████████▌ | 693/809 [13:35<02:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141     0.4423     0.1233        282        640:  86%|████████▌ | 695/809 [13:37<02:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.142     0.4422     0.1234        471        640:  86%|████████▌ | 697/809 [13:40<02:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141     0.4424     0.1234        160        640:  86%|████████▋ | 699/809 [13:42<02:06,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141     0.4424     0.1234        319        640:  87%|████████▋ | 700/809 [13:43<02:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.142     0.4423     0.1235        198        640:  87%|████████▋ | 702/809 [13:45<02:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.142     0.4421     0.1237        509        640:  87%|████████▋ | 704/809 [13:48<02:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.142     0.4422     0.1236        387        640:  87%|████████▋ | 705/809 [13:49<02:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.143     0.4422     0.1237        278        640:  87%|████████▋ | 707/809 [13:51<01:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.143     0.4423     0.1236        258        640:  88%|████████▊ | 709/809 [13:54<01:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.143     0.4425     0.1236        102        640:  88%|████████▊ | 710/809 [13:55<01:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.144     0.4423      0.124        235        640:  88%|████████▊ | 712/809 [13:57<01:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.142     0.4425     0.1238        130        640:  88%|████████▊ | 714/809 [14:00<01:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141     0.4428     0.1237        229        640:  89%|████████▊ | 716/809 [14:02<01:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141     0.4428     0.1236        217        640:  89%|████████▊ | 717/809 [14:03<01:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141      0.443     0.1235        332        640:  89%|████████▉ | 719/809 [14:06<01:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141      0.443     0.1234        246        640:  89%|████████▉ | 721/809 [14:08<01:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141      0.443     0.1234        234        640:  89%|████████▉ | 723/809 [14:10<01:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141     0.4428     0.1235        303        640:  90%|████████▉ | 725/809 [14:13<01:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141     0.4429     0.1234        166        640:  90%|████████▉ | 726/809 [14:14<01:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141      0.443     0.1233        211        640:  90%|████████▉ | 728/809 [14:16<01:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141     0.4431     0.1232        166        640:  90%|█████████ | 730/809 [14:18<01:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.14     0.4432     0.1232        114        640:  90%|█████████ | 731/809 [14:20<01:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.141     0.4431     0.1232        261        640:  91%|█████████ | 733/809 [14:22<01:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.14     0.4434     0.1232        287        640:  91%|█████████ | 735/809 [14:24<01:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.14     0.4437      0.123        176        640:  91%|█████████ | 737/809 [14:27<01:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.14     0.4437      0.123        355        640:  91%|█████████ | 738/809 [14:28<01:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.139     0.4441     0.1229        136        640:  91%|█████████▏| 740/809 [14:30<01:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.139     0.4442      0.123        350        640:  92%|█████████▏| 742/809 [14:33<01:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.138     0.4444      0.123        107        640:  92%|█████████▏| 743/809 [14:34<01:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.138     0.4446     0.1229        211        640:  92%|█████████▏| 745/809 [14:36<01:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.137     0.4449     0.1228        190        640:  92%|█████████▏| 747/809 [14:38<01:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.137      0.445     0.1227        117        640:  92%|█████████▏| 748/809 [14:40<01:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.136     0.4451     0.1227        284        640:  93%|█████████▎| 750/809 [14:42<01:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.136     0.4454     0.1226        306        640:  93%|█████████▎| 752/809 [14:44<01:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.136     0.4453     0.1228        353        640:  93%|█████████▎| 754/809 [14:47<01:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.136     0.4454     0.1227        144        640:  93%|█████████▎| 755/809 [14:48<01:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.136     0.4454     0.1227        385        640:  94%|█████████▎| 757/809 [14:50<01:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.135     0.4457     0.1226        100        640:  94%|█████████▍| 759/809 [14:53<00:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.136     0.4456     0.1227        284        640:  94%|█████████▍| 760/809 [14:54<00:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.135     0.4457     0.1226        161        640:  94%|█████████▍| 762/809 [14:56<00:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.135     0.4458     0.1225        254        640:  94%|█████████▍| 764/809 [14:58<00:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.134      0.446     0.1224        145        640:  95%|█████████▍| 766/809 [15:01<00:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.134     0.4461     0.1224        210        640:  95%|█████████▍| 767/809 [15:02<00:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.133     0.4462     0.1222        245        640:  95%|█████████▌| 769/809 [15:04<00:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.133     0.4464     0.1223         82        640:  95%|█████████▌| 771/809 [15:06<00:43,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.132     0.4466     0.1222        287        640:  96%|█████████▌| 773/809 [15:09<00:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.132     0.4465     0.1222        330        640:  96%|█████████▌| 774/809 [15:10<00:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4469     0.1221        219        640:  96%|█████████▌| 776/809 [15:12<00:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.13     0.4471     0.1221         47        640:  96%|█████████▌| 778/809 [15:15<00:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.13     0.4472      0.122        127        640:  96%|█████████▋| 779/809 [15:16<00:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.13     0.4473     0.1219        148        640:  97%|█████████▋| 781/809 [15:18<00:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.13     0.4472     0.1219        540        640:  97%|█████████▋| 783/809 [15:20<00:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.129     0.4475     0.1218         99        640:  97%|█████████▋| 785/809 [15:23<00:27,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.129     0.4475     0.1218        269        640:  97%|█████████▋| 786/809 [15:24<00:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.128     0.4475     0.1217        204        640:  97%|█████████▋| 788/809 [15:26<00:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.128     0.4474     0.1216        235        640:  98%|█████████▊| 790/809 [15:29<00:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.129     0.4473     0.1217        407        640:  98%|█████████▊| 791/809 [15:30<00:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.13      0.447      0.122        623        640:  98%|█████████▊| 793/809 [15:32<00:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4468     0.1221        322        640:  98%|█████████▊| 795/809 [15:34<00:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4469     0.1221        357        640:  99%|█████████▊| 797/809 [15:37<00:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4469     0.1221        395        640:  99%|█████████▊| 798/809 [15:38<00:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4469     0.1222        503        640:  99%|█████████▉| 800/809 [15:40<00:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4469     0.1221        327        640:  99%|█████████▉| 802/809 [15:43<00:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4469     0.1221        213        640:  99%|█████████▉| 803/809 [15:44<00:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G       1.13     0.4469     0.1221        550        640: 100%|█████████▉| 805/809 [15:46<00:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4468      0.122        288        640: 100%|█████████▉| 807/809 [15:49<00:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     72/300      13.5G      1.131     0.4469      0.122        186        640: 100%|█████████▉| 808/809 [15:50<00:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  10%|█         | 7/69 [00:01<00:14,  4.21it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  22%|██▏       | 15/69 [00:03<00:13,  4.10it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  33%|███▎      | 23/69 [00:05<00:11,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  45%|████▍     | 31/69 [00:07<00:09,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  58%|███���█▊    | 40/69 [00:09<00:07,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  68%|██████▊   | 47/69 [00:11<00:06,  3.54it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  80%|███████▉  | 55/69 [00:13<00:03,  3.93it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  91%|█████████▏| 63/69 [00:15<00:01,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  91%|█████████▏| 63/69 [00:15<00:01,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.04it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.04it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300      13.9G      1.414     0.3984     0.2752        281        640:   0%|          | 1/809 [00:01<16:30,  1.23s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300      13.9G      1.226     0.4289     0.1802        143        640:   0%|          | 2/809 [00:02<15:55,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300      13.9G      1.251     0.4268     0.1485        249        640:   0%|          | 4/809 [00:04<15:24,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.319     0.4003     0.1449        294        640:   1%|          | 6/809 [00:06<15:33,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.296     0.4031      0.131        340        640:   1%|          | 8/809 [00:09<15:23,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.293     0.4037     0.1381        302        640:   1%|          | 9/809 [00:10<15:20,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.292     0.3992     0.1336        346        640:   1%|▏         | 11/809 [00:12<15:29,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.266     0.4057     0.1294        353        640:   2%|▏         | 13/809 [00:15<15:27,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.247     0.4102     0.1312        476        640:   2%|▏         | 15/809 [00:17<15:18,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.243     0.4097     0.1286        180        640:   2%|▏         | 16/809 [00:18<15:16,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.225     0.4178      0.123        235        640:   2%|▏         | 18/809 [00:20<15:19,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.214     0.4217      0.124        223        640:   2%|▏         | 20/809 [00:23<15:22,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.224      0.418     0.1238        337        640:   3%|▎         | 21/809 [00:24<15:33,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.214     0.4227      0.122        371        640:   3%|▎         | 23/809 [00:26<15:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.198     0.4269     0.1189        172        640:   3%|▎         | 25/809 [00:29<15:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.189     0.4293      0.117        286        640:   3%|▎         | 26/809 [00:30<15:17,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.167     0.4365     0.1146        221        640:   3%|▎         | 28/809 [00:32<15:24,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.16     0.4394      0.115        140        640:   4%|▎         | 30/809 [00:35<15:29,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.173     0.4368     0.1185        464        640:   4%|▍         | 31/809 [00:36<15:25,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.184     0.4326     0.1193        440        640:   4%|▍         | 33/809 [00:38<15:03,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.192     0.4288     0.1208        600        640:   4%|▍         | 35/809 [00:40<15:12,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.176     0.4343     0.1184        123        640:   5%|▍         | 37/809 [00:43<15:07,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.173     0.4351     0.1177        158        640:   5%|▍         | 38/809 [00:44<15:04,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.167     0.4361     0.1186        222        640:   5%|▍         | 40/809 [00:46<14:48,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.169     0.4363     0.1205        329        640:   5%|▌         | 42/809 [00:49<14:46,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.167     0.4371     0.1201        306        640:   5%|▌         | 44/809 [00:51<14:57,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.168     0.4383     0.1206        410        640:   6%|▌         | 45/809 [00:52<14:59,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.153     0.4416     0.1185        200        640:   6%|▌         | 47/809 [00:54<14:53,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.14     0.4456     0.1168        113        640:   6%|▌         | 49/809 [00:57<14:45,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.144     0.4445     0.1182        384        640:   6%|▌         | 50/809 [00:58<14:45,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.152      0.442     0.1189        250        640:   6%|▋         | 52/809 [01:00<14:38,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.153     0.4421     0.1195        195        640:   7%|▋         | 54/809 [01:03<14:51,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.158     0.4402     0.1201        417        640:   7%|▋         | 55/809 [01:04<15:00,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.153     0.4412     0.1188        182        640:   7%|▋         | 57/809 [01:06<15:22,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.152     0.4431     0.1191        100        640:   7%|▋         | 59/809 [01:09<14:53,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.148     0.4442      0.118        232        640:   8%|▊         | 61/809 [01:11<14:38,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.154     0.4423     0.1178        288        640:   8%|▊         | 63/809 [01:13<14:19,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.156      0.442      0.118        256        640:   8%|▊         | 64/809 [01:14<14:13,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.155     0.4428     0.1174        328        640:   8%|▊         | 66/809 [01:17<14:18,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.16      0.441     0.1173        353        640:   8%|▊         | 68/809 [01:19<14:05,  1.14s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.16     0.4415     0.1176        491        640:   9%|▊         | 70/809 [01:21<14:15,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.158     0.4421     0.1178        214        640:   9%|▉         | 71/809 [01:23<14:08,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.152     0.4438     0.1174        188        640:   9%|▉         | 73/809 [01:25<14:00,  1.14s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.15     0.4435     0.1174        490        640:   9%|▉         | 75/809 [01:27<14:13,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.148      0.444     0.1166        299        640:  10%|▉         | 77/809 [01:30<14:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.144     0.4446     0.1162        214        640:  10%|▉         | 78/809 [01:31<14:16,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.142     0.4456     0.1171        469        640:  10%|▉         | 80/809 [01:33<14:27,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.138     0.4468     0.1165        109        640:  10%|█         | 82/809 [01:35<14:09,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.139     0.4468      0.117        318        640:  10%|█         | 83/809 [01:37<14:33,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.141     0.4463     0.1169        203        640:  11%|█         | 85/809 [01:39<14:24,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.132     0.4491     0.1161        177        640:  11%|█         | 87/809 [01:41<14:09,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.129     0.4501     0.1156        165        640:  11%|█         | 88/809 [01:43<14:00,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.126     0.4507     0.1153        115        640:  11%|█         | 90/809 [01:45<14:10,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.12      0.452     0.1148        240        640:  11%|█▏        | 92/809 [01:47<13:57,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.117     0.4526     0.1147        318        640:  12%|█▏        | 94/809 [01:50<14:06,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.116      0.453     0.1146        212        640:  12%|█▏        | 95/809 [01:51<14:03,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.113     0.4542     0.1146        166        640:  12%|█▏        | 97/809 [01:53<13:58,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4554     0.1141        202        640:  12%|█▏        | 99/809 [01:55<13:44,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4545     0.1149        283        640:  12%|█▏        | 100/809 [01:57<13:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4552     0.1142        363        640:  13%|█▎        | 102/809 [01:59<13:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4571     0.1135        170        640:  13%|█▎        | 104/809 [02:01<13:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4568     0.1139        171        640:  13%|█▎        | 106/809 [02:04<13:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4576     0.1137        170        640:  13%|█▎        | 107/809 [02:05<13:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4571     0.1137        307        640:  13%|█▎        | 109/809 [02:07<14:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.097     0.4584     0.1131        218        640:  14%|█▎        | 111/809 [02:10<13:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.095     0.4593     0.1127         96        640:  14%|█▍        | 112/809 [02:11<13:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.09     0.4601     0.1121        233        640:  14%|█▍        | 114/809 [02:13<13:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.087     0.4617     0.1116        158        640:  14%|█▍        | 116/809 [02:16<13:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.086     0.4619     0.1115        291        640:  14%|█▍        | 117/809 [02:17<13:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.089     0.4604     0.1112        277        640:  15%|█▍        | 119/809 [02:19<13:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.092     0.4591     0.1119        229        640:  15%|█▍        | 121/809 [02:21<13:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.089     0.4597     0.1115        171        640:  15%|█▌        | 122/809 [02:23<13:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.086     0.4604     0.1115         64        640:  15%|█▌        | 124/809 [02:25<13:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.085     0.4606      0.111        131        640:  16%|█▌        | 126/809 [02:27<13:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.082     0.4618     0.1108        115        640:  16%|█▌        | 128/809 [02:30<13:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.084     0.4613     0.1111        368        640:  16%|█▌        | 129/809 [02:31<13:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.085     0.4612      0.112        475        640:  16%|█▌        | 131/809 [02:33<13:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.086     0.4604      0.112        418        640:  16%|█▋        | 133/809 [02:36<13:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.086     0.4606     0.1118        173        640:  17%|█▋        | 134/809 [02:37<13:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.087     0.4605     0.1119        108        640:  17%|█▋        | 136/809 [02:39<12:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.087     0.4607     0.1117         57        640:  17%|█▋        | 138/809 [02:41<13:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.09     0.4606     0.1116        171        640:  17%|█▋        | 140/809 [02:44<12:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.09     0.4607     0.1115        331        640:  17%|█▋        | 141/809 [02:45<13:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.093     0.4601     0.1121        572        640:  18%|█▊        | 143/809 [02:47<13:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.092     0.4605     0.1118        178        640:  18%|█▊        | 145/809 [02:49<12:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.091     0.4609     0.1114         66        640:  18%|█▊        | 147/809 [02:52<12:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.088     0.4616     0.1112        116        640:  18%|█▊        | 148/809 [02:53<12:39,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.088     0.4619     0.1111        228        640:  19%|█▊        | 150/809 [02:55<12:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.087     0.4621     0.1107        240        640:  19%|█▉        | 152/809 [02:58<12:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.087      0.462     0.1107        346        640:  19%|█▉        | 153/809 [02:59<12:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.086     0.4623     0.1104        281        640:  19%|█▉        | 155/809 [03:01<12:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.085     0.4624     0.1103        223        640:  19%|█▉        | 157/809 [03:03<12:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.087     0.4614     0.1103        165        640:  20%|█▉        | 159/809 [03:06<12:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.093     0.4599     0.1111        589        640:  20%|█▉        | 160/809 [03:07<12:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.099     0.4578     0.1121        430        640:  20%|██        | 162/809 [03:09<12:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4561     0.1124        352        640:  20%|██        | 164/809 [03:12<12:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4555     0.1124        333        640:  20%|██        | 165/809 [03:13<12:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4556     0.1127        120        640:  21%|██        | 167/809 [03:15<12:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4546      0.113        369        640:  21%|██        | 169/809 [03:18<12:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4544      0.113        180        640:  21%|██        | 171/809 [03:20<12:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11      0.454     0.1129        295        640:  21%|██▏       | 172/809 [03:21<12:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4549     0.1127        194        640:  22%|██▏       | 174/809 [03:23<12:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4553     0.1123        132        640:  22%|██▏       | 176/809 [03:26<12:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105      0.456     0.1119        104        640:  22%|██▏       | 177/809 [03:27<12:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4559     0.1117        364        640:  22%|██▏       | 179/809 [03:29<12:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4557     0.1117        380        640:  22%|██▏       | 181/809 [03:32<12:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4557     0.1115        294        640:  23%|██▎       | 183/809 [03:34<12:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4556     0.1116        256        640:  23%|██▎       | 184/809 [03:35<12:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4555     0.1114        280        640:  23%|██▎       | 186/809 [03:37<12:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105      0.456     0.1112        173        640:  23%|██▎       | 188/809 [03:40<11:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4553     0.1112        395        640:  23%|██▎       | 189/809 [03:41<12:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4551     0.1114        180        640:  24%|██▎       | 191/809 [03:43<12:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4544     0.1115        403        640:  24%|██▍       | 193/809 [03:46<12:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4545     0.1114        185        640:  24%|██▍       | 195/809 [03:48<11:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4545     0.1112        206        640:  24%|██▍       | 196/809 [03:49<11:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109      0.454     0.1114        357        640:  24%|██▍       | 198/809 [03:52<12:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4531     0.1113        229        640:  25%|██▍       | 200/809 [03:54<11:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.113     0.4524     0.1114        338        640:  25%|██▍       | 201/809 [03:55<11:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.114     0.4518     0.1113        208        640:  25%|██▌       | 203/809 [03:57<11:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.115     0.4515     0.1111        178        640:  25%|██▌       | 205/809 [04:00<11:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.116     0.4513     0.1115        383        640:  25%|██▌       | 206/809 [04:01<11:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4518     0.1111        222        640:  26%|██▌       | 208/809 [04:03<11:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4515     0.1111        359        640:  26%|██▌       | 210/809 [04:06<11:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4523     0.1107        115        640:  26%|██▌       | 212/809 [04:08<11:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111      0.452     0.1108        234        640:  26%|██▋       | 214/809 [04:10<11:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111      0.452     0.1106         94        640:  27%|██▋       | 216/809 [04:12<11:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.113     0.4514      0.111        369        640:  27%|██▋       | 217/809 [04:14<11:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4516     0.1109        184        640:  27%|██▋       | 219/809 [04:16<11:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4516      0.111        392        640:  27%|██▋       | 221/809 [04:18<11:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.113     0.4512      0.111        225        640:  27%|██▋       | 222/809 [04:20<11:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.117     0.4501     0.1111        155        640:  28%|██▊       | 224/809 [04:22<11:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.121     0.4489     0.1121        187        640:  28%|██▊       | 226/809 [04:24<11:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.123      0.448     0.1124        546        640:  28%|██▊       | 227/809 [04:26<11:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.127     0.4468     0.1127        277        640:  28%|██▊       | 229/809 [04:28<11:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.127     0.4467     0.1127        231        640:  29%|██▊       | 231/809 [04:30<11:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.128     0.4461     0.1126        187        640:  29%|██▊       | 232/809 [04:31<11:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.128     0.4465     0.1125        129        640:  29%|██▉       | 234/809 [04:34<11:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.127     0.4467     0.1123        112        640:  29%|██▉       | 236/809 [04:36<10:57,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.125     0.4472     0.1122         98        640:  29%|██▉       | 238/809 [04:38<10:55,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.123     0.4475      0.112        240        640:  30%|██▉       | 239/809 [04:40<10:52,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.122     0.4483     0.1118        180        640:  30%|██▉       | 241/809 [04:42<10:48,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.122     0.4484     0.1117        285        640:  30%|███       | 243/809 [04:44<10:46,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.121     0.4489     0.1115        194        640:  30%|███       | 245/809 [04:47<11:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.121     0.4487     0.1115        159        640:  30%|███       | 246/809 [04:48<11:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.12     0.4488     0.1113        228        640:  31%|███       | 248/809 [04:50<10:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.118     0.4494     0.1112        255        640:  31%|███       | 250/809 [04:52<10:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.116     0.4499     0.1109        108        640:  31%|███       | 252/809 [04:55<10:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.116       0.45     0.1108        302        640:  31%|███▏      | 253/809 [04:56<11:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.115     0.4502     0.1107        167        640:  32%|███▏      | 255/809 [04:58<10:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.113     0.4508     0.1104        180        640:  32%|███▏      | 257/809 [05:01<11:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.113     0.4508     0.1105        278        640:  32%|███▏      | 258/809 [05:02<10:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.113     0.4508     0.1104        338        640:  32%|███▏      | 260/809 [05:04<10:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4511     0.1102        203        640:  32%|███▏      | 262/809 [05:07<10:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4512     0.1101        249        640:  33%|███▎      | 263/809 [05:08<10:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4519     0.1098        129        640:  33%|███▎      | 265/809 [05:10<10:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4517       0.11        491        640:  33%|███▎      | 267/809 [05:13<10:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.113     0.4511     0.1112        754        640:  33%|███▎      | 268/809 [05:14<11:00,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.114     0.4505     0.1114        275        640:  33%|███▎      | 270/809 [05:16<10:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.114     0.4503     0.1114        514        640:  34%|███▎      | 272/809 [05:19<10:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4512     0.1113        319        640:  34%|███▎      | 273/809 [05:20<10:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4513     0.1112        449        640:  34%|███▍      | 275/809 [05:22<10:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4519      0.111        294        640:  34%|███▍      | 277/809 [05:24<10:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4515     0.1109        451        640:  34%|███▍      | 278/809 [05:26<10:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.114     0.4505     0.1125        431        640:  35%|███▍      | 280/809 [05:28<10:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.115     0.4502     0.1127        369        640:  35%|███▍      | 282/809 [05:31<10:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.114     0.4506     0.1127        119        640:  35%|███▍      | 283/809 [05:32<10:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4511     0.1125        234        640:  35%|███▌      | 285/809 [05:34<10:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4513     0.1123        131        640:  35%|███▌      | 287/809 [05:36<10:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4513     0.1126        211        640:  36%|███▌      | 289/809 [05:39<10:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4513     0.1125        104        640:  36%|███▌      | 290/809 [05:40<10:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4511     0.1128        268        640:  36%|███▌      | 292/809 [05:42<09:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4511     0.1127        222        640:  36%|███▋      | 294/809 [05:45<09:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4512     0.1127        202        640:  36%|███▋      | 295/809 [05:46<09:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4513     0.1127        202        640:  37%|███▋      | 297/809 [05:48<09:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4509     0.1127        333        640:  37%|███▋      | 299/809 [05:50<09:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.113     0.4507     0.1125        226        640:  37%|███▋      | 301/809 [05:53<10:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4508     0.1125        240        640:  37%|███▋      | 302/809 [05:54<10:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4514     0.1124        164        640:  38%|███▊      | 304/809 [05:56<09:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4513     0.1123        268        640:  38%|███▊      | 306/809 [05:59<09:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4511     0.1122        226        640:  38%|███▊      | 307/809 [06:00<09:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4515      0.112        140        640:  38%|███▊      | 309/809 [06:02<09:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4518     0.1118        469        640:  38%|███▊      | 311/809 [06:04<09:34,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4513     0.1124        339        640:  39%|███▊      | 313/809 [06:07<09:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4513     0.1123        266        640:  39%|███▉      | 314/809 [06:08<09:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4514     0.1126        250        640:  39%|███▉      | 316/809 [06:10<09:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4519     0.1124        197        640:  39%|███▉      | 318/809 [06:13<09:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4521     0.1122        231        640:  40%|███▉      | 320/809 [06:15<09:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4521     0.1122        391        640:  40%|███▉      | 321/809 [06:16<09:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4524      0.112         81        640:  40%|███▉      | 323/809 [06:18<09:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4528     0.1118        171        640:  40%|████      | 325/809 [06:21<09:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102      0.453     0.1116        177        640:  40%|████      | 326/809 [06:22<09:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4533     0.1115        148        640:  41%|████      | 328/809 [06:24<09:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4533     0.1114        265        640:  41%|████      | 330/809 [06:27<09:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4532     0.1118        267        640:  41%|████      | 331/809 [06:28<09:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.099     0.4538     0.1116        185        640:  41%|████      | 333/809 [06:30<09:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4534     0.1116        169        640:  41%|████▏     | 335/809 [06:33<09:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4529     0.1115        101        640:  42%|████▏     | 337/809 [06:35<09:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4528     0.1114        170        640:  42%|████▏     | 338/809 [06:36<09:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4524     0.1114        276        640:  42%|████▏     | 340/809 [06:38<09:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4515     0.1116        236        640:  42%|████▏     | 342/809 [06:41<09:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4514     0.1116        421        640:  42%|████▏     | 343/809 [06:42<09:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4518     0.1116        175        640:  43%|████▎     | 345/809 [06:44<08:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4522     0.1114        119        640:  43%|████▎     | 347/809 [06:47<08:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4522     0.1115        462        640:  43%|████▎     | 349/809 [06:49<09:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4522     0.1115         90        640:  43%|████▎     | 350/809 [06:50<08:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4524     0.1113        175        640:  44%|████▎     | 352/809 [06:52<08:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4526     0.1111        190        640:  44%|████▍     | 354/809 [06:55<08:40,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4524     0.1111        171        640:  44%|████▍     | 356/809 [06:57<08:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4523     0.1111        282        640:  44%|████▍     | 357/809 [06:58<08:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4521      0.111        189        640:  44%|████▍     | 359/809 [07:01<08:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101      0.452     0.1108        341        640:  45%|████▍     | 361/809 [07:03<08:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101      0.452     0.1108        172        640:  45%|████▍     | 362/809 [07:04<08:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4521     0.1108        155        640:  45%|████▍     | 364/809 [07:06<08:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4521     0.1108        182        640:  45%|████▌     | 366/809 [07:09<08:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1      0.452     0.1113        180        640:  45%|████▌     | 368/809 [07:11<08:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4519     0.1114        230        640:  46%|████▌     | 370/809 [07:13<08:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.099     0.4524     0.1113        130        640:  46%|████▌     | 371/809 [07:15<08:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4524     0.1113        188        640:  46%|████▌     | 373/809 [07:17<08:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4523     0.1114        634        640:  46%|████▋     | 375/809 [07:19<08:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4524     0.1115        266        640:  47%|████▋     | 377/809 [07:21<08:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4525     0.1115        163        640:  47%|████▋     | 378/809 [07:23<08:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4523     0.1114        130        640:  47%|████▋     | 380/809 [07:25<08:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4525     0.1113        209        640:  47%|████▋     | 382/809 [07:27<08:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4525     0.1112        244        640:  47%|████▋     | 383/809 [07:29<08:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4522     0.1118        335        640:  48%|████▊     | 385/809 [07:31<08:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4522     0.1117        287        640:  48%|████▊     | 387/809 [07:33<08:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4526     0.1123        478        640:  48%|████▊     | 389/809 [07:36<08:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4528     0.1122        183        640:  48%|████▊     | 390/809 [07:37<08:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4527     0.1123        176        640:  48%|████▊     | 392/809 [07:39<08:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4523     0.1124        299        640:  49%|████▊     | 394/809 [07:41<08:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103      0.452     0.1124        190        640:  49%|████▉     | 396/809 [07:44<07:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4519     0.1123        402        640:  49%|████▉     | 397/809 [07:45<08:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4516     0.1125        395        640:  49%|████▉     | 399/809 [07:47<07:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4519     0.1124         85        640:  50%|████▉     | 401/809 [07:50<07:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4517     0.1126        348        640:  50%|████▉     | 402/809 [07:51<07:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4519     0.1127        392        640:  50%|████▉     | 404/809 [07:53<07:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103      0.452     0.1126        384        640:  50%|█████     | 406/809 [07:55<07:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4524     0.1124        172        640:  50%|█████     | 408/809 [07:58<07:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4523     0.1123        276        640:  51%|█████     | 409/809 [07:59<07:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4523     0.1124        354        640:  51%|█████     | 411/809 [08:01<07:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4526     0.1122        154        640:  51%|█████     | 413/809 [08:04<07:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4526     0.1122        294        640:  51%|█████     | 414/809 [08:05<07:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4529      0.112        253        640:  51%|█████▏    | 416/809 [08:07<07:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4531     0.1119        217        640:  52%|█████▏    | 418/809 [08:09<07:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.099     0.4534     0.1118        243        640:  52%|█████▏    | 420/809 [08:12<07:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.099     0.4534     0.1119        369        640:  52%|█████▏    | 421/809 [08:13<07:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.099     0.4533     0.1119        257        640:  52%|█████▏    | 423/809 [08:15<07:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1      0.453      0.112        123        640:  53%|█████▎    | 425/809 [08:18<07:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G        1.1     0.4529      0.112        138        640:  53%|█████▎    | 426/809 [08:19<07:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4522     0.1124        462        640:  53%|█████▎    | 428/809 [08:21<07:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4522     0.1123        343        640:  53%|█████▎    | 430/809 [08:24<07:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101     0.4528     0.1122        173        640:  53%|█████▎    | 432/809 [08:26<07:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4526     0.1122        316        640:  54%|█████▎    | 433/809 [08:27<07:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4527     0.1121        205        640:  54%|█████▍    | 435/809 [08:29<07:09,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.101      0.453     0.1121        238        640:  54%|█████▍    | 437/809 [08:32<07:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4529     0.1123        315        640:  54%|█████▍    | 438/809 [08:33<07:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.102     0.4527     0.1124        246        640:  54%|█████▍    | 440/809 [08:35<07:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4527     0.1126        458        640:  55%|█████▍    | 442/809 [08:37<07:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4527     0.1128        100        640:  55%|█████▍    | 444/809 [08:40<07:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4526     0.1128        273        640:  55%|█████▌    | 445/809 [08:41<07:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4523     0.1127        264        640:  55%|█████▌    | 447/809 [08:43<07:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105      0.452     0.1127        219        640:  56%|█████▌    | 449/809 [08:46<07:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4518     0.1127        315        640:  56%|█████▌    | 450/809 [08:47<07:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4516     0.1127        161        640:  56%|█████▌    | 452/809 [08:49<06:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4514     0.1126        233        640:  56%|█████▌    | 454/809 [08:52<07:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4516     0.1125        130        640:  56%|█████▌    | 455/809 [08:53<06:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4516     0.1125        406        640:  56%|█████▋    | 457/809 [08:55<06:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4515      0.113        245        640:  57%|█████▋    | 459/809 [08:57<06:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4516      0.113        361        640:  57%|█████▋    | 461/809 [09:00<06:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4516      0.113        187        640:  57%|█████▋    | 462/809 [09:01<06:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4519     0.1128        116        640:  57%|█████▋    | 464/809 [09:04<06:59,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4518     0.1129         93        640:  58%|█████▊    | 466/809 [09:06<06:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4517     0.1129        278        640:  58%|█████▊    | 467/809 [09:07<06:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4521     0.1127         59        640:  58%|█████▊    | 469/809 [09:09<06:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4522     0.1128        381        640:  58%|█████▊    | 471/809 [09:12<06:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4521     0.1128        247        640:  58%|█████▊    | 472/809 [09:13<06:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4519     0.1131        343        640:  59%|█████▊    | 474/809 [09:15<06:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107      0.452     0.1129        160        640:  59%|█████▉    | 476/809 [09:18<06:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4518     0.1129        287        640:  59%|█████▉    | 478/809 [09:20<06:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4516     0.1128        264        640:  59%|█████▉    | 479/809 [09:21<06:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4515     0.1128        176        640:  59%|█████▉    | 481/809 [09:24<06:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4512     0.1128        187        640:  60%|█████▉    | 483/809 [09:26<06:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109      0.451     0.1128        171        640:  60%|█████▉    | 484/809 [09:27<06:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4509      0.113        119        640:  60%|██████    | 486/809 [09:29<06:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4509     0.1131        102        640:  60%|██████    | 488/809 [09:32<06:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4509     0.1131        228        640:  60%|██████    | 489/809 [09:33<06:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4506     0.1132        330        640:  61%|██████    | 491/809 [09:35<06:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4508      0.113        163        640:  61%|██████    | 493/809 [09:38<06:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4509     0.1129        328        640:  61%|██████    | 495/809 [09:40<06:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4509     0.1129        215        640:  61%|██████▏   | 496/809 [09:41<06:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4508      0.113         85        640:  62%|██████▏   | 498/809 [09:43<05:55,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4507      0.113        255        640:  62%|██████▏   | 500/809 [09:46<05:55,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4508      0.113        286        640:  62%|██████▏   | 502/809 [09:48<05:51,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4507      0.113        431        640:  62%|██████▏   | 503/809 [09:49<05:50,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4509     0.1129        261        640:  62%|██████▏   | 505/809 [09:51<05:48,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4508     0.1129        328        640:  63%|██████▎   | 507/809 [09:54<05:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4505     0.1131        293        640:  63%|██████▎   | 509/809 [09:56<05:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4507      0.113        182        640:  63%|██████▎   | 510/809 [09:57<05:44,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4508     0.1129        250        640:  63%|██████▎   | 512/809 [09:59<05:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4514     0.1128        144        640:  64%|██████▎   | 514/809 [10:02<05:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4513     0.1128        323        640:  64%|██████▎   | 515/809 [10:03<06:01,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4516     0.1127        285        640:  64%|██████▍   | 517/809 [10:06<06:14,  1.28s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4516     0.1127        162        640:  64%|██████▍   | 518/809 [10:07<06:14,  1.29s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4518     0.1126        112        640:  64%|██████▍   | 520/809 [10:09<05:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4519      0.113        181        640:  65%|██████▍   | 522/809 [10:12<05:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4516     0.1132        269        640:  65%|██████▍   | 524/809 [10:14<05:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4515     0.1132        387        640:  65%|██████▍   | 525/809 [10:15<05:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4515     0.1132        285        640:  65%|██████▌   | 527/809 [10:18<05:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4516     0.1131        338        640:  65%|██████▌   | 529/809 [10:20<05:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4516      0.113        302        640:  66%|██████▌   | 531/809 [10:23<05:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4516     0.1131        205        640:  66%|██████▌   | 532/809 [10:24<05:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4516      0.113        435        640:  66%|██████▌   | 534/809 [10:26<05:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104      0.452      0.113        170        640:  66%|█��████▋   | 536/809 [10:28<05:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4521     0.1131        332        640:  67%|██████▋   | 538/809 [10:31<05:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4522      0.113        178        640:  67%|██████▋   | 539/809 [10:32<05:11,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4525     0.1129        197        640:  67%|██████▋   | 541/809 [10:34<05:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4523     0.1131        398        640:  67%|██████▋   | 543/809 [10:37<05:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4524     0.1131        249        640:  67%|██████▋   | 544/809 [10:38<05:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.103     0.4525      0.113        177        640:  67%|██████▋   | 546/809 [10:40<05:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4523     0.1132        513        640:  68%|██████▊   | 548/809 [10:42<05:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4521     0.1133        382        640:  68%|██████▊   | 549/809 [10:44<05:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4518     0.1135        318        640:  68%|██████▊   | 551/809 [10:46<05:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4517     0.1134        297        640:  68%|██████▊   | 553/809 [10:48<04:55,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4519     0.1133        289        640:  69%|██████▊   | 555/809 [10:50<04:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4519     0.1133        239        640:  69%|██████▊   | 556/809 [10:52<04:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4519     0.1133        267        640:  69%|██████▉   | 558/809 [10:54<04:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106      0.452     0.1132        220        640:  69%|██████▉   | 560/809 [10:57<05:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4522     0.1132        185        640:  69%|██████▉   | 561/809 [10:58<04:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4525     0.1131        272        640:  70%|██████▉   | 563/809 [11:00<04:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4525     0.1131        175        640:  70%|██████▉   | 565/809 [11:03<04:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4525     0.1131        242        640:  70%|██████▉   | 566/809 [11:04<04:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4521     0.1133        198        640:  70%|███████   | 568/809 [11:06<04:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4518     0.1135        315        640:  70%|███████   | 570/809 [11:08<04:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4518     0.1136        225        640:  71%|███████   | 572/809 [11:11<04:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4518     0.1137        343        640:  71%|███████   | 573/809 [11:12<04:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4519     0.1136        218        640:  71%|███████   | 575/809 [11:14<04:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4518     0.1137        477        640:  71%|███████▏  | 577/809 [11:16<04:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4518     0.1136        219        640:  71%|███████▏  | 578/809 [11:18<04:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4521     0.1136        103        640:  72%|███████▏  | 580/809 [11:20<04:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4518     0.1141        202        640:  72%|███████▏  | 582/809 [11:23<04:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4517      0.114        209        640:  72%|███████▏  | 583/809 [11:24<04:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4514     0.1141        291        640:  72%|███████▏  | 585/809 [11:26<04:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4513     0.1141        267        640:  73%|███████▎  | 587/809 [11:28<04:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4511     0.1141        378        640:  73%|███████▎  | 589/809 [11:31<04:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4512     0.1141        214        640:  73%|███████▎  | 590/809 [11:32<04:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4511     0.1141        223        640:  73%|███████▎  | 592/809 [11:34<04:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11      0.451     0.1142        266        640:  73%|███████▎  | 594/809 [11:37<04:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4512     0.1142        205        640:  74%|███████▎  | 595/809 [11:38<04:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4515     0.1141        120        640:  74%|███████▍  | 597/809 [11:40<04:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4513     0.1142        202        640:  74%|███████▍  | 599/809 [11:43<04:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4512     0.1142        189        640:  74%|███████▍  | 600/809 [11:44<04:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4512     0.1143        130        640:  74%|███████▍  | 602/809 [11:46<04:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4511     0.1142        388        640:  75%|███████▍  | 604/809 [11:49<03:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4512     0.1142        185        640:  75%|███████▍  | 606/809 [11:51<03:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4513     0.1142        149        640:  75%|███████▌  | 607/809 [11:52<03:52,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4512     0.1142        135        640:  75%|███████▌  | 609/809 [11:54<03:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4513     0.1142        301        640:  76%|███████▌  | 611/809 [11:57<03:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4515     0.1141        175        640:  76%|███████▌  | 612/809 [11:58<03:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4517     0.1143        377        640:  76%|███████▌  | 614/809 [12:00<03:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4514     0.1144        307        640:  76%|███████▌  | 616/809 [12:02<03:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4515     0.1144        317        640:  76%|███████▋  | 618/809 [12:05<03:39,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4513     0.1144        517        640:  77%|███████▋  | 619/809 [12:06<03:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109      0.451     0.1144        263        640:  77%|███████▋  | 621/809 [12:08<03:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109      0.451     0.1144        339        640:  77%|███████▋  | 623/809 [12:11<03:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109      0.451     0.1143        180        640:  77%|███████▋  | 625/809 [12:13<03:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4509     0.1153        494        640:  77%|███████▋  | 626/809 [12:14<03:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109      0.451     0.1152        108        640:  78%|███████▊  | 628/809 [12:17<03:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4512     0.1151        244        640:  78%|███████▊  | 630/809 [12:19<03:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4512      0.115        113        640:  78%|███████▊  | 631/809 [12:20<03:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4514     0.1149        282        640:  78%|███████▊  | 633/809 [12:22<03:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4515     0.1148        370        640:  78%|███████▊  | 635/809 [12:25<03:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4515     0.1147        150        640:  79%|███████▊  | 636/809 [12:26<03:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4518     0.1146        137        640:  79%|███████▉  | 638/809 [12:28<03:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105      0.452     0.1146        178        640:  79%|███████▉  | 640/809 [12:31<03:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105      0.452     0.1145        144        640:  79%|███████▉  | 642/809 [12:33<03:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4521     0.1145        207        640:  79%|███████▉  | 643/809 [12:34<03:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4521     0.1145        350        640:  80%|███████▉  | 645/809 [12:37<03:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4522     0.1144        266        640:  80%|███████▉  | 647/809 [12:39<03:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4524     0.1144        208        640:  80%|████████  | 648/809 [12:40<03:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4521     0.1144        449        640:  80%|████████  | 650/809 [12:42<03:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4523     0.1143        164        640:  81%|████████  | 652/809 [12:45<03:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4522     0.1143        274        640:  81%|████████  | 653/809 [12:46<03:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4518     0.1144        449        640:  81%|████████  | 655/809 [12:48<03:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4516     0.1146        242        640:  81%|████████  | 657/809 [12:51<02:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4517     0.1147        155        640:  81%|████████▏ | 659/809 [12:53<02:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4518     0.1147        265        640:  82%|████████▏ | 660/809 [12:54<03:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105      0.452     0.1147        388        640:  82%|████████▏ | 662/809 [12:57<02:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4522     0.1146        139        640:  82%|████████▏ | 663/809 [12:58<02:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4522     0.1146        255        640:  82%|████████▏ | 665/809 [13:00<02:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4523     0.1146        247        640:  82%|████████▏ | 667/809 [13:03<02:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4524     0.1145        200        640:  83%|████████▎ | 668/809 [13:04<02:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4525     0.1146        116        640:  83%|████████▎ | 670/809 [13:06<02:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4523     0.1148        287        640:  83%|████████▎ | 672/809 [13:09<02:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4522     0.1148        409        640:  83%|████████▎ | 674/809 [13:11<02:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4523     0.1148        130        640:  83%|████████▎ | 675/809 [13:12<02:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.104     0.4521      0.115        537        640:  84%|████████▎ | 677/809 [13:15<02:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.105     0.4518     0.1151        351        640:  84%|████████▍ | 679/809 [13:17<02:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4514     0.1152        622        640:  84%|████████▍ | 681/809 [13:19<02:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4513     0.1152        269        640:  84%|████████▍ | 683/809 [13:22<02:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4514     0.1152        213        640:  85%|████████▍ | 684/809 [13:23<02:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4512     0.1152        133        640:  85%|████████▍ | 686/809 [13:25<02:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4512     0.1151        174        640:  85%|████████▌ | 688/809 [13:27<02:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4513     0.1151        425        640:  85%|████████▌ | 689/809 [13:29<02:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4514      0.115        109        640:  85%|████████▌ | 691/809 [13:31<02:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.106     0.4512     0.1151        317        640:  86%|████████▌ | 693/809 [13:33<02:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107      0.451     0.1151        260        640:  86%|████████▌ | 695/809 [13:36<02:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.107     0.4509     0.1151        255        640:  86%|████████▌ | 696/809 [13:37<02:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4507     0.1151        408        640:  86%|████████▋ | 698/809 [13:39<02:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4505     0.1152        488        640:  87%|████████▋ | 700/809 [13:41<02:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4503     0.1152        290        640:  87%|████████▋ | 701/809 [13:43<02:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4502     0.1152        317        640:  87%|████████▋ | 703/809 [13:45<02:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4503     0.1151        179        640:  87%|████████▋ | 705/809 [13:47<02:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4502     0.1151        300        640:  87%|████████▋ | 707/809 [13:50<02:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4501     0.1151        356        640:  88%|████████▊ | 708/809 [13:51<01:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4499     0.1152        230        640:  88%|████████▊ | 710/809 [13:53<01:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109       0.45     0.1151        127        640:  88%|████████▊ | 712/809 [13:56<01:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109       0.45     0.1151        278        640:  88%|████████▊ | 713/809 [13:57<01:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109       0.45      0.115        284        640:  88%|████████▊ | 715/809 [13:59<01:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.108     0.4501      0.115        243        640:  89%|████████▊ | 717/809 [14:01<01:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109       0.45      0.115        304        640:  89%|████████▉ | 719/809 [14:04<01:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4501     0.1149        239        640:  89%|████████▉ | 720/809 [14:05<01:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109       0.45     0.1149        227        640:  89%|████████▉ | 722/809 [14:07<01:40,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.109     0.4497     0.1149        424        640:  89%|████████▉ | 724/809 [14:10<01:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.11     0.4495     0.1149        228        640:  90%|████████▉ | 725/809 [14:11<01:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4492      0.115        369        640:  90%|████████▉ | 727/809 [14:13<01:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.111     0.4492      0.115        174        640:  90%|█████████ | 729/809 [14:15<01:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4489     0.1154        376        640:  90%|█████████ | 731/809 [14:18<01:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.113     0.4488     0.1155        396        640:  90%|█████████ | 732/809 [14:19<01:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.113     0.4487     0.1155        273        640:  91%|█████████ | 734/809 [14:21<01:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112      0.449     0.1155        112        640:  91%|█████████ | 736/809 [14:24<01:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112      0.449     0.1155        275        640:  91%|█████████ | 737/809 [14:25<01:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112      0.449     0.1154        193        640:  91%|█████████▏| 739/809 [14:27<01:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4491     0.1154        239        640:  92%|█████████▏| 741/809 [14:30<01:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4491     0.1153        249        640:  92%|█████████▏| 743/809 [14:32<01:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4491     0.1153        339        640:  92%|█████████▏| 744/809 [14:33<01:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4491     0.1153        349        640:  92%|█████████▏| 746/809 [14:35<01:12,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112      0.449     0.1152        243        640:  92%|█████████▏| 748/809 [14:38<01:10,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4489     0.1153        321        640:  93%|█████████▎| 749/809 [14:39<01:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4489     0.1153        351        640:  93%|█████████▎| 751/809 [14:41<01:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4488     0.1153        178        640:  93%|█████████▎| 753/809 [14:43<01:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.112     0.4487     0.1154        447        640:  93%|█████████▎| 755/809 [14:46<01:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.114     0.4485     0.1165        452        640:  93%|█████████▎| 756/809 [14:47<01:05,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.114     0.4485     0.1164        156        640:  94%|█████████▎| 758/809 [14:50<01:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.115     0.4482     0.1164        265        640:  94%|█████████▍| 760/809 [14:52<00:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.114     0.4482     0.1164        199        640:  94%|█████████▍| 761/809 [14:53<00:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.115     0.4481     0.1166        112        640:  94%|█████████▍| 763/809 [14:55<00:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.117     0.4477     0.1176        556        640:  94%|█████████▍| 764/809 [14:57<00:56,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.117     0.4476     0.1176        258        640:  95%|█████████▍| 766/809 [14:59<00:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.117     0.4477     0.1176        397        640:  95%|█████████▍| 768/809 [15:02<00:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.117     0.4477     0.1176        270        640:  95%|█████████▌| 769/809 [15:03<00:48,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.117     0.4478     0.1176        225        640:  95%|█████████▌| 771/809 [15:05<00:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.117     0.4477     0.1176        154        640:  96%|█████████▌| 773/809 [15:08<00:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.118     0.4474     0.1178        394        640:  96%|█████████▌| 775/809 [15:10<00:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.118     0.4473     0.1178        267        640:  96%|█████████▌| 776/809 [15:11<00:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.118     0.4472     0.1177        186        640:  96%|█████████▌| 778/809 [15:13<00:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.119     0.4469      0.118        505        640:  96%|█████████▋| 780/809 [15:16<00:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.12     0.4467     0.1179        208        640:  97%|█████████▋| 781/809 [15:17<00:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.12     0.4468     0.1179        334        640:  97%|█████████▋| 783/809 [15:19<00:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.119     0.4469      0.118        203        640:  97%|█████████▋| 785/809 [15:22<00:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.12     0.4467     0.1182        332        640:  97%|█████████▋| 787/809 [15:24<00:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.12     0.4466     0.1182        263        640:  97%|█████████▋| 788/809 [15:25<00:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.12     0.4466     0.1181        306        640:  98%|█████████▊| 790/809 [15:28<00:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G       1.12     0.4466     0.1181        532        640:  98%|█████████▊| 792/809 [15:30<00:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.121     0.4464     0.1182        337        640:  98%|█████████▊| 793/809 [15:31<00:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.121     0.4464     0.1183        399        640:  98%|█████████▊| 795/809 [15:33<00:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.121     0.4463     0.1183        249        640:  99%|█████████▊| 797/809 [15:36<00:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.121     0.4463     0.1183        125        640:  99%|█████████▊| 798/809 [15:37<00:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.122      0.446     0.1184        292        640:  99%|█████████▉| 800/809 [15:39<00:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.122     0.4457     0.1184        374        640:  99%|█████████▉| 802/809 [15:42<00:08,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.122     0.4457     0.1184        392        640:  99%|█████████▉| 804/809 [15:44<00:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.122     0.4456     0.1184        324        640: 100%|█████████▉| 805/809 [15:45<00:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     73/300        14G      1.122     0.4456     0.1184        190        640: 100%|█████████▉| 807/809 [15:47<00:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   1%|▏         | 1/69 [00:00<00:17,  3.86it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  14%|█▍        | 10/69 [00:02<00:14,  4.16it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  26%|██▌       | 18/69 [00:04<00:12,  4.12it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  38%|███▊      | 26/69 [00:06<00:10,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  49%|████▉     | 34/69 [00:08<00:09,  3.87it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  61%|██████    | 42/69 [00:10<00:06,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  71%|███████   | 49/69 [00:12<00:05,  3.76it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  83%|████████▎ | 57/69 [00:14<00:03,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:00,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:00,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G     0.9436     0.5179      0.108        203        640:   0%|          | 2/809 [00:02<15:12,  1.13s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      0.895     0.5335     0.1032        118        640:   0%|          | 3/809 [00:03<15:12,  1.13s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G     0.9632     0.5027     0.1085        157        640:   1%|          | 5/809 [00:05<15:21,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.087     0.4542     0.1177        323        640:   1%|          | 7/809 [00:08<15:37,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.063     0.4619     0.1157        256        640:   1%|          | 8/809 [00:09<15:33,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.06     0.4624     0.1159        250        640:   1%|          | 10/809 [00:11<15:25,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.115     0.4471     0.1343        120        640:   1%|▏         | 12/809 [00:13<15:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.089     0.4507     0.1295        243        640:   2%|▏         | 14/809 [00:16<15:31,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.08     0.4534     0.1276        165        640:   2%|▏         | 15/809 [00:17<15:23,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.097     0.4506     0.1258        204        640:   2%|▏         | 17/809 [00:19<15:31,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.108     0.4496     0.1316        398        640:   2%|▏         | 19/809 [00:22<15:44,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.124     0.4459     0.1346        440        640:   2%|▏         | 20/809 [00:23<15:54,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.117     0.4478     0.1328        221        640:   3%|▎         | 22/809 [00:25<15:32,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.099     0.4553     0.1297        140        640:   3%|▎         | 24/809 [00:28<15:32,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.106      0.453      0.134        324        640:   3%|▎         | 25/809 [00:29<16:11,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.108     0.4518     0.1333        389        640:   3%|▎         | 27/809 [00:31<15:50,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.104     0.4532     0.1319        330        640:   4%|▎         | 29/809 [00:34<15:31,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.114     0.4503     0.1332        479        640:   4%|▎         | 30/809 [00:35<15:23,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.113     0.4502      0.132        376        640:   4%|▍         | 32/809 [00:37<15:12,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.123     0.4477     0.1333        394        640:   4%|▍         | 34/809 [00:40<15:50,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.119      0.449     0.1322        282        640:   4%|▍         | 35/809 [00:41<15:28,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.112     0.4499     0.1307        200        640:   5%|▍         | 37/809 [00:43<15:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.115     0.4468     0.1305        378        640:   5%|▍         | 39/809 [00:46<15:16,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.111     0.4497     0.1299        140        640:   5%|▍         | 40/809 [00:47<15:15,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.102     0.4514     0.1286        255        640:   5%|▌         | 42/809 [00:49<14:50,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.101     0.4519     0.1267        315        640:   5%|▌         | 44/809 [00:52<15:23,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.097     0.4522     0.1262        428        640:   6%|▌         | 46/809 [00:54<15:14,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.089     0.4545      0.125        183        640:   6%|▌         | 48/809 [00:56<14:59,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.085     0.4557     0.1247        205        640:   6%|▌         | 49/809 [00:58<14:57,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.085      0.456     0.1235        180        640:   6%|▋         | 51/809 [01:00<14:45,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G        1.1     0.4511     0.1275        149        640:   7%|▋         | 53/809 [01:02<14:46,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.097     0.4521     0.1273        191        640:   7%|▋         | 54/809 [01:03<14:37,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G        1.1     0.4516     0.1281        307        640:   7%|▋         | 56/809 [01:06<14:28,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.097     0.4524     0.1278        278        640:   7%|▋         | 58/809 [01:08<14:26,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.094     0.4514     0.1268        448        640:   7%|▋         | 60/809 [01:10<14:37,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.096     0.4509     0.1265        241        640:   8%|▊         | 61/809 [01:11<14:27,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.109     0.4464      0.127        377        640:   8%|▊         | 63/809 [01:14<14:31,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.107     0.4463     0.1259        224        640:   8%|▊         | 65/809 [01:16<14:19,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.104     0.4462     0.1252        145        640:   8%|▊         | 67/809 [01:18<14:15,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.101     0.4471     0.1248        328        640:   8%|▊         | 68/809 [01:20<14:24,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.107     0.4459     0.1274        193        640:   9%|▊         | 70/809 [01:22<14:17,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G        1.1     0.4472     0.1264        108        640:   9%|▉         | 72/809 [01:24<14:16,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G        1.1     0.4466     0.1261        340        640:   9%|▉         | 73/809 [01:25<14:22,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.106     0.4451       0.13        131        640:   9%|▉         | 75/809 [01:28<14:25,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.107     0.4444       0.13        416        640:  10%|▉         | 77/809 [01:30<14:36,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.11     0.4437       0.13        418        640:  10%|▉         | 78/809 [01:32<14:43,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.11     0.4436     0.1297        198        640:  10%|▉         | 80/809 [01:34<14:30,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.112     0.4435     0.1293        373        640:  10%|█         | 82/809 [01:36<14:28,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.114     0.4425     0.1291        324        640:  10%|█         | 83/809 [01:37<14:22,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.121     0.4406     0.1326        352        640:  11%|█         | 85/809 [01:40<14:33,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.124     0.4401     0.1327        222        640:  11%|█         | 87/809 [01:42<14:31,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.123     0.4405     0.1323        347        640:  11%|█         | 88/809 [01:43<14:14,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.134     0.4369     0.1344        473        640:  11%|█         | 90/809 [01:46<14:20,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.137     0.4359     0.1341        283        640:  11%|█▏        | 92/809 [01:48<14:18,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.138     0.4355     0.1338        317        640:  11%|█▏        | 93/809 [01:49<14:15,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.141     0.4348     0.1332        190        640:  12%|█▏        | 95/809 [01:52<14:12,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4334     0.1337        247        640:  12%|█▏        | 97/809 [01:54<14:09,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146      0.434     0.1334        300        640:  12%|█▏        | 98/809 [01:55<14:08,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.152     0.4316     0.1334        364        640:  12%|█▏        | 100/809 [01:58<14:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.153     0.4318     0.1327        192        640:  13%|█▎        | 102/809 [02:00<14:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4333     0.1319        165        640:  13%|█▎        | 104/809 [02:03<13:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4335     0.1316        299        640:  13%|█▎        | 105/809 [02:04<13:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.145     0.4355      0.131        123        640:  13%|█▎        | 107/809 [02:06<13:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4356     0.1304        351        640:  13%|█▎        | 109/809 [02:08<13:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4355     0.1302        343        640:  14%|█▎        | 110/809 [02:10<13:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4345       0.13        372        640:  14%|█▍        | 112/809 [02:12<13:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4361     0.1292        209        640:  14%|█▍        | 114/809 [02:14<13:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.145     0.4357     0.1291        287        640:  14%|█▍        | 115/809 [02:15<13:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4348     0.1289        280        640:  14%|█▍        | 117/809 [02:18<13:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.153      0.434     0.1288        387        640:  15%|█▍        | 119/809 [02:20<13:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.154     0.4337     0.1286        158        640:  15%|█▍        | 120/809 [02:22<13:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.159     0.4321     0.1289        491        640:  15%|█▌        | 122/809 [02:24<13:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.165     0.4304     0.1289        438        640:  15%|█▌        | 124/809 [02:26<13:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.164     0.4306     0.1288        309        640:  15%|█▌        | 125/809 [02:28<13:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.161     0.4321     0.1279        295        640:  16%|█▌        | 127/809 [02:30<13:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.161     0.4321     0.1275        202        640:  16%|█▌        | 129/809 [02:32<13:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.161     0.4321     0.1279        231        640:  16%|█▌        | 131/809 [02:35<13:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.162     0.4325     0.1278        148        640:  16%|█▋        | 132/809 [02:36<13:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.165     0.4317     0.1282        409        640:  17%|█▋        | 134/809 [02:38<13:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.168      0.431     0.1288        333        640:  17%|█▋        | 136/809 [02:40<13:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.167     0.4321     0.1289        165        640:  17%|█▋        | 138/809 [02:43<12:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.167     0.4326     0.1288        138        640:  17%|█▋        | 139/809 [02:44<13:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.168     0.4326     0.1287        361        640:  17%|█▋        | 141/809 [02:46<13:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.165     0.4333     0.1282        185        640:  18%|█▊        | 143/809 [02:49<12:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.164     0.4336     0.1278         96        640:  18%|█▊        | 144/809 [02:50<12:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.164     0.4336     0.1276        245        640:  18%|█▊        | 146/809 [02:52<12:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.159     0.4345     0.1269        194        640:  18%|█▊        | 148/809 [02:54<12:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.161     0.4338     0.1271        689        640:  19%|█▊        | 150/809 [02:57<12:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.164     0.4328     0.1273        259        640:  19%|█▊        | 151/809 [02:58<12:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.163     0.4328     0.1272        184        640:  19%|█▉        | 153/809 [03:00<12:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.167     0.4316     0.1275        215        640:  19%|█▉        | 155/809 [03:03<12:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.166     0.4316     0.1274        252        640:  19%|█▉        | 156/809 [03:04<12:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.167     0.4318     0.1272        400        640:  20%|█▉        | 158/809 [03:06<12:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.166     0.4322      0.127        225        640:  20%|█▉        | 160/809 [03:09<12:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.166     0.4325     0.1269        183        640:  20%|█▉        | 161/809 [03:10<12:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4319     0.1285        340        640:  20%|██        | 163/809 [03:12<12:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.169     0.4318     0.1286        183        640:  20%|██        | 165/809 [03:15<12:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172      0.431     0.1288        210        640:  21%|██        | 167/809 [03:17<12:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4315     0.1284        228        640:  21%|██        | 168/809 [03:18<12:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4313     0.1282        230        640:  21%|██        | 170/809 [03:20<12:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4308     0.1283        235        640:  21%|██▏       | 172/809 [03:23<12:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.173     0.4308     0.1283        169        640:  21%|██▏       | 173/809 [03:24<12:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.177     0.4297     0.1292        331        640:  22%|██▏       | 175/809 [03:26<12:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.174     0.4314     0.1288        142        640:  22%|██▏       | 177/809 [03:28<12:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4323     0.1284        123        640:  22%|██▏       | 179/809 [03:31<12:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4323     0.1283        178        640:  22%|██▏       | 180/809 [03:32<12:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.174     0.4314     0.1285        222        640:  22%|██▏       | 182/809 [03:34<12:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.175     0.4312     0.1288        191        640:  23%|██▎       | 184/809 [03:37<12:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.175     0.4314     0.1287        186        640:  23%|██▎       | 185/809 [03:38<12:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.176     0.4308     0.1286        281        640:  23%|██▎       | 187/809 [03:40<11:55,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.177     0.4302     0.1285        237        640:  23%|██▎       | 189/809 [03:42<11:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.177     0.4304     0.1284        246        640:  24%|██▎       | 191/809 [03:45<11:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.177     0.4305     0.1282        205        640:  24%|██▎       | 192/809 [03:46<11:52,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.176     0.4302     0.1281        305        640:  24%|██▍       | 194/809 [03:48<11:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.175     0.4306     0.1281        146        640:  24%|██▍       | 196/809 [03:51<11:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.176     0.4306      0.128        414        640:  24%|██▍       | 197/809 [03:52<11:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4299     0.1287        290        640:  25%|██▍       | 199/809 [03:54<11:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.178       0.43     0.1286        196        640:  25%|██▍       | 201/809 [03:56<11:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4298     0.1284        197        640:  25%|██▌       | 203/809 [03:59<11:39,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.178     0.4301     0.1282        229        640:  25%|██▌       | 205/809 [04:01<11:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4295     0.1284        410        640:  26%|██▌       | 207/809 [04:03<11:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4291     0.1285        374        640:  26%|██▌       | 208/809 [04:05<11:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4294     0.1285        154        640:  26%|██▌       | 210/809 [04:07<11:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.181     0.4289     0.1286        139        640:  26%|██▌       | 212/809 [04:09<11:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4291     0.1285        329        640:  26%|██▋       | 213/809 [04:10<11:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.178     0.4299     0.1282         88        640:  27%|██▋       | 215/809 [04:13<11:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.178     0.4299      0.128        190        640:  27%|██▋       | 217/809 [04:15<11:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4293     0.1288        247        640:  27%|██▋       | 219/809 [04:17<11:18,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.182     0.4289     0.1292        443        640:  27%|██▋       | 220/809 [04:19<11:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.185     0.4281     0.1299        193        640:  27%|██▋       | 222/809 [04:21<11:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.184     0.4287     0.1296        216        640:  28%|██▊       | 224/809 [04:23<11:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.183      0.429     0.1294        163        640:  28%|██▊       | 226/809 [04:26<11:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.184     0.4288     0.1294        364        640:  28%|██▊       | 227/809 [04:27<11:10,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.183     0.4292     0.1292        304        640:  28%|██▊       | 229/809 [04:29<11:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.182     0.4296      0.129        122        640:  29%|██▊       | 231/809 [04:31<11:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.181     0.4297     0.1288        248        640:  29%|██▊       | 232/809 [04:33<11:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4302     0.1284        341        640:  29%|██▉       | 234/809 [04:35<11:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4304     0.1285        277        640:  29%|██▉       | 236/809 [04:37<11:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4305     0.1284        210        640:  29%|██▉       | 237/809 [04:39<11:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.181       0.43     0.1286        433        640:  30%|██▉       | 239/809 [04:41<11:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4304     0.1283        273        640:  30%|██▉       | 241/809 [04:43<11:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4303     0.1283        389        640:  30%|███       | 243/809 [04:46<10:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4301     0.1282        348        640:  30%|███       | 244/809 [04:47<11:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4298     0.1281        407        640:  30%|███       | 246/809 [04:49<11:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4303     0.1278        344        640:  31%|███       | 248/809 [04:51<10:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4299      0.128        492        640:  31%|███       | 249/809 [04:53<10:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.181     0.4299      0.128        385        640:  31%|███       | 251/809 [04:55<10:43,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4299     0.1278        210        640:  31%|███▏      | 253/809 [04:57<10:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4299     0.1278        127        640:  32%|███▏      | 255/809 [05:00<10:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4302     0.1279        119        640:  32%|███▏      | 256/809 [05:01<10:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4298     0.1283        387        640:  32%|███▏      | 258/809 [05:03<10:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.182      0.429     0.1282        410        640:  32%|███▏      | 260/809 [05:06<10:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4296      0.128         86        640:  32%|███▏      | 261/809 [05:07<10:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4295     0.1278        309        640:  33%|███▎      | 263/809 [05:09<10:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.177       0.43     0.1275        122        640:  33%|███▎      | 265/809 [05:11<10:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.177       0.43     0.1275        299        640:  33%|███▎      | 266/809 [05:13<10:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4292     0.1282        405        640:  33%|███▎      | 268/809 [05:15<10:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4292     0.1281        223        640:  33%|███▎      | 270/809 [05:17<10:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4296     0.1281        310        640:  34%|███▎      | 272/809 [05:20<10:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.177     0.4304      0.128        111        640:  34%|███▎      | 273/809 [05:21<10:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.177     0.4303     0.1279        310        640:  34%|███▍      | 275/809 [05:23<10:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.177     0.4304      0.128        471        640:  34%|███▍      | 277/809 [05:26<10:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4299     0.1287        419        640:  34%|███▍      | 278/809 [05:27<10:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4299     0.1287        321        640:  35%|███▍      | 280/809 [05:29<10:10,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4296     0.1291        297        640:  35%|███▍      | 282/809 [05:31<10:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4301      0.129        255        640:  35%|███▌      | 284/809 [05:34<10:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4298     0.1292        279        640:  35%|███▌      | 285/809 [05:35<10:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.178     0.4302      0.129        239        640:  35%|███▌      | 287/809 [05:37<10:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4298     0.1298        170        640:  36%|███▌      | 289/809 [05:40<10:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.18     0.4294     0.1298        329        640:  36%|███▌      | 290/809 [05:41<10:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.178     0.4301     0.1297        213        640:  36%|███▌      | 292/809 [05:43<10:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4299     0.1295        182        640:  36%|███▋      | 294/809 [05:46<10:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.181     0.4292     0.1297        409        640:  36%|███▋      | 295/809 [05:47<09:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.184     0.4282     0.1299        403        640:  37%|███▋      | 297/809 [05:49<09:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.185      0.428     0.1298        162        640:  37%|███▋      | 299/809 [05:51<09:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.186     0.4277     0.1298        293        640:  37%|███▋      | 300/809 [05:53<10:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.186     0.4277     0.1297        167        640:  37%|███▋      | 302/809 [05:55<09:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.187     0.4273     0.1304        532        640:  38%|███▊      | 304/809 [05:57<10:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.187     0.4272     0.1302        325        640:  38%|███▊      | 306/809 [06:00<10:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.187     0.4273     0.1302        159        640:  38%|███▊      | 307/809 [06:01<09:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.187     0.4275     0.1299        215        640:  38%|███▊      | 309/809 [06:03<09:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.187     0.4275     0.1299        499        640:  38%|███▊      | 311/809 [06:06<09:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.187     0.4274     0.1298        241        640:  39%|███▊      | 312/809 [06:07<09:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.187     0.4273     0.1296        239        640:  39%|███▉      | 314/809 [06:09<09:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.187     0.4275     0.1294        168        640:  39%|███▉      | 316/809 [06:12<10:20,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.185     0.4277     0.1292         96        640:  39%|███▉      | 317/809 [06:13<10:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.186     0.4277     0.1293        306        640:  39%|███▉      | 319/809 [06:15<09:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.186     0.4276     0.1293        430        640:  40%|███▉      | 321/809 [06:18<09:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.187     0.4274     0.1292        392        640:  40%|███▉      | 322/809 [06:19<09:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.187     0.4274     0.1293        209        640:  40%|████      | 324/809 [06:21<09:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.189     0.4267     0.1295        358        640:  40%|████      | 326/809 [06:24<09:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4266     0.1293        170        640:  41%|████      | 328/809 [06:26<09:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4264     0.1294        244        640:  41%|████      | 329/809 [06:27<09:12,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.189     0.4268     0.1292        169        640:  41%|████      | 331/809 [06:29<09:07,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.191     0.4264     0.1297        171        640:  41%|████      | 333/809 [06:32<09:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.192     0.4262     0.1296        315        640:  41%|████▏     | 335/809 [06:34<09:04,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.193     0.4259     0.1296        492        640:  42%|████▏     | 336/809 [06:35<09:02,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.195     0.4253     0.1297        366        640:  42%|████▏     | 338/809 [06:37<09:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.196     0.4251     0.1296        261        640:  42%|████▏     | 340/809 [06:40<09:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.194     0.4255     0.1294        145        640:  42%|████▏     | 341/809 [06:41<09:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.193     0.4258     0.1291        193        640:  42%|████▏     | 343/809 [06:43<09:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.193     0.4262     0.1289        211        640:  43%|████▎     | 345/809 [06:46<09:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.191     0.4266     0.1288        236        640:  43%|████▎     | 347/809 [06:48<08:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.192     0.4265     0.1287        376        640:  43%|████▎     | 349/809 [06:50<08:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.192     0.4264     0.1286        311        640:  43%|████▎     | 350/809 [06:51<08:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.193     0.4261     0.1285        200        640:  44%|████▎     | 352/809 [06:54<08:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.192     0.4264     0.1283        145        640:  44%|████▍     | 354/809 [06:56<08:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4268      0.128        155        640:  44%|████▍     | 356/809 [06:58<08:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4266      0.128        467        640:  44%|████▍     | 357/809 [07:00<08:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4265      0.128        421        640:  44%|████▍     | 359/809 [07:02<08:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4264     0.1279        322        640:  45%|████▍     | 361/809 [07:04<08:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.191     0.4262      0.128        312        640:  45%|████▍     | 362/809 [07:05<08:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.191     0.4261     0.1279        310        640:  45%|████▍     | 364/809 [07:08<08:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4264     0.1277        168        640:  45%|████▌     | 366/809 [07:10<08:31,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.189     0.4265     0.1275        366        640:  45%|████▌     | 368/809 [07:12<08:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4264     0.1275        191        640:  46%|████▌     | 369/809 [07:14<08:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.189     0.4265     0.1273        350        640:  46%|████▌     | 371/809 [07:16<08:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4263     0.1276        317        640:  46%|████▌     | 373/809 [07:18<08:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.192      0.426     0.1283        369        640:  46%|████▌     | 374/809 [07:19<08:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.192     0.4259     0.1283        446        640:  46%|████▋     | 376/809 [07:22<08:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.193     0.4256     0.1285        221        640:  47%|████▋     | 378/809 [07:24<08:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.195     0.4251     0.1286        479        640:  47%|████▋     | 380/809 [07:26<08:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.195     0.4252     0.1285        184        640:  47%|████▋     | 381/809 [07:28<08:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.194     0.4256     0.1283        257        640:  47%|████▋     | 383/809 [07:30<08:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.193     0.4259     0.1282        157        640:  48%|████▊     | 385/809 [07:32<08:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.194     0.4257     0.1292        418        640:  48%|████▊     | 386/809 [07:34<08:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.192     0.4265      0.129        253        640:  48%|████▊     | 388/809 [07:36<08:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4271     0.1288        196        640:  48%|████▊     | 390/809 [07:38<08:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4271     0.1289        191        640:  48%|████▊     | 392/809 [07:41<08:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19      0.427     0.1288        234        640:  49%|████▊     | 393/809 [07:42<08:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.189     0.4273     0.1286        216        640:  49%|████▉     | 395/809 [07:44<07:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.19     0.4272     0.1286        244        640:  49%|████▉     | 397/809 [07:47<08:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.188     0.4276     0.1284        168        640:  49%|████▉     | 398/809 [07:48<08:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.185     0.4285     0.1281         64        640:  49%|████▉     | 400/809 [07:50<08:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.184     0.4288     0.1279        215        640:  50%|████▉     | 402/809 [07:52<07:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.184     0.4289     0.1278        255        640:  50%|████▉     | 403/809 [07:53<07:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.184      0.429     0.1276        286        640:  50%|█████     | 405/809 [07:56<07:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.182       0.43     0.1274        205        640:  50%|█████     | 407/809 [07:58<07:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.181     0.4302     0.1272        170        640:  51%|█████     | 409/809 [08:00<07:39,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.182       0.43     0.1276        390        640:  51%|█████     | 410/809 [08:02<07:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.181     0.4301     0.1274        278        640:  51%|█████     | 412/809 [08:04<07:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.179     0.4307     0.1272        185        640:  51%|█████     | 414/809 [08:06<07:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.178     0.4311     0.1269        152        640:  51%|█████▏    | 416/809 [08:09<07:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.177     0.4313     0.1268        182        640:  52%|█████▏    | 417/809 [08:10<07:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.177     0.4314     0.1269        160        640:  52%|█████▏    | 419/809 [08:12<07:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.176     0.4315     0.1266        249        640:  52%|█████▏    | 421/809 [08:14<07:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.176     0.4315     0.1265        240        640:  52%|█████▏    | 422/809 [08:16<07:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.175     0.4317     0.1264        174        640:  52%|█████▏    | 424/809 [08:18<07:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.174     0.4322     0.1262        119        640:  53%|█████▎    | 426/809 [08:20<07:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.174     0.4321     0.1261        381        640:  53%|█████▎    | 427/809 [08:21<07:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4326     0.1259        245        640:  53%|█████▎    | 429/809 [08:24<07:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4328     0.1259        267        640:  53%|█████▎    | 431/809 [08:26<07:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4329     0.1257        362        640:  54%|█████▎    | 433/809 [08:29<07:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4328     0.1256        352        640:  54%|█████▎    | 434/809 [08:30<07:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4327     0.1256        239        640:  54%|█████▍    | 436/809 [08:32<07:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4327     0.1255        241        640:  54%|█████▍    | 438/809 [08:34<07:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4327     0.1255        196        640:  54%|█████▍    | 439/809 [08:36<07:06,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4327     0.1254        187        640:  55%|█████▍    | 441/809 [08:38<07:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4328     0.1252        194        640:  55%|█████▍    | 443/809 [08:40<07:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4327     0.1252        226        640:  55%|█████▌    | 445/809 [08:43<07:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171      0.433      0.125        123        640:  55%|█████▌    | 446/809 [08:44<07:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4334     0.1249        389        640:  55%|█████▌    | 448/809 [08:46<06:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4333     0.1249        129        640:  56%|█████▌    | 450/809 [08:48<06:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4333     0.1251        472        640:  56%|█████▌    | 451/809 [08:50<07:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4332     0.1251        277        640:  56%|█████▌    | 453/809 [08:52<07:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4335     0.1251        125        640:  56%|█████▌    | 455/809 [08:54<06:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4336     0.1252        433        640:  56%|█████▋    | 457/809 [08:57<06:45,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4339     0.1251         86        640:  57%|█████▋    | 458/809 [08:58<06:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4337     0.1254        257        640:  57%|█████▋    | 460/809 [09:00<06:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4335     0.1255        214        640:  57%|█████▋    | 462/809 [09:03<06:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4334     0.1254        207        640:  57%|█████▋    | 463/809 [09:04<06:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4332     0.1254        312        640:  57%|█████▋    | 465/809 [09:06<06:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4336     0.1253        140        640:  58%|█████▊    | 467/809 [09:08<06:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4338     0.1252        172        640:  58%|█████▊    | 468/809 [09:10<06:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.169     0.4342     0.1251        143        640:  58%|█████▊    | 470/809 [09:12<06:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17      0.434     0.1252        673        640:  58%|█████▊    | 472/809 [09:14<06:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4339     0.1251        315        640:  59%|█████▊    | 474/809 [09:17<06:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4338     0.1251        296        640:  59%|█████▊    | 475/809 [09:18<06:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4335     0.1251        304        640:  59%|█████▉    | 477/809 [09:20<06:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4333      0.125        253        640:  59%|█████▉    | 479/809 [09:23<06:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4333      0.125        147        640:  59%|█████▉    | 480/809 [09:24<06:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4331     0.1252        186        640:  60%|█████▉    | 482/809 [09:26<06:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4331     0.1253        405        640:  60%|█████▉    | 484/809 [09:29<06:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4334     0.1252         97        640:  60%|█████▉    | 485/809 [09:30<06:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.172     0.4335     0.1253        235        640:  60%|██████    | 487/809 [09:32<06:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4335     0.1252        195        640:  60%|██████    | 489/809 [09:34<06:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4336     0.1251        198        640:  61%|██████    | 491/809 [09:37<06:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4338      0.125        276        640:  61%|██████    | 492/809 [09:38<06:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4337     0.1251        445        640:  61%|██████    | 494/809 [09:40<06:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.171     0.4337     0.1251        614        640:  61%|██████▏   | 496/809 [09:43<06:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.17     0.4341      0.125        204        640:  62%|██████▏   | 498/809 [09:45<06:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.169     0.4343     0.1249        104        640:  62%|██████▏   | 499/809 [09:46<06:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.169      0.434      0.125        363        640:  62%|██████▏   | 501/809 [09:48<05:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.169     0.4343     0.1249        136        640:  62%|██████▏   | 503/809 [09:51<05:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.167     0.4346     0.1247        207        640:  62%|██████▏   | 505/809 [09:53<05:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.167     0.4348     0.1246        166        640:  63%|██████▎   | 506/809 [09:54<05:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.165     0.4354     0.1244        179        640:  63%|██████▎   | 508/809 [09:57<06:17,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.164     0.4357     0.1242        184        640:  63%|██████▎   | 510/809 [09:59<05:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.164     0.4356     0.1242        232        640:  63%|██████▎   | 511/809 [10:00<05:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.164     0.4355     0.1241        143        640:  63%|██████▎   | 513/809 [10:03<05:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.163     0.4357     0.1239        206        640:  64%|██████▎   | 515/809 [10:05<05:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.163     0.4358     0.1239        273        640:  64%|██████▍   | 516/809 [10:06<05:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.162     0.4362     0.1238        163        640:  64%|██████▍   | 518/809 [10:09<05:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.162     0.4361     0.1237        255        640:  64%|██████▍   | 520/809 [10:11<05:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.162     0.4362     0.1236        269        640:  65%|██████▍   | 522/809 [10:14<05:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.163     0.4361     0.1237        432        640:  65%|██████▍   | 523/809 [10:15<05:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.163      0.436     0.1237        372        640:  65%|██████▍   | 525/809 [10:17<05:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.164     0.4358     0.1238        216        640:  65%|██████▌   | 527/809 [10:20<05:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.163     0.4361     0.1237        133        640:  65%|██████▌   | 528/809 [10:21<05:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.162     0.4367     0.1235        120        640:  66%|██████▌   | 530/809 [10:23<05:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.162     0.4368     0.1235        299        640:  66%|██████▌   | 532/809 [10:25<05:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.161     0.4369     0.1234        246        640:  66%|██████▌   | 533/809 [10:27<05:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.16     0.4373     0.1232        188        640:  66%|██████▌   | 535/809 [10:29<05:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.161     0.4372     0.1236        162        640:  66%|██████▋   | 537/809 [10:31<05:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.16     0.4374     0.1235        152        640:  67%|██████▋   | 538/809 [10:32<05:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.16     0.4374     0.1237        197        640:  67%|██████▋   | 540/809 [10:35<05:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.159     0.4379     0.1236        105        640:  67%|██████▋   | 542/809 [10:37<05:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.159     0.4379     0.1237        171        640:  67%|██████▋   | 544/809 [10:39<05:04,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.158      0.438     0.1236        335        640:  67%|██████▋   | 545/809 [10:40<05:02,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.159     0.4378     0.1238        561        640:  68%|██████▊   | 547/809 [10:43<05:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.159     0.4378     0.1237        283        640:  68%|██████▊   | 549/809 [10:45<05:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.159     0.4379     0.1236        153        640:  68%|██████▊   | 550/809 [10:47<05:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.159      0.438     0.1235        137        640:  68%|██████▊   | 552/809 [10:49<05:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.159     0.4379     0.1238        477        640:  68%|██████▊   | 554/809 [10:51<04:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.16     0.4379     0.1238        477        640:  69%|██████▊   | 556/809 [10:54<05:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.16     0.4377     0.1238        352        640:  69%|██████▉   | 557/809 [10:55<04:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.16     0.4378     0.1237        201        640:  69%|██████▉   | 559/809 [10:57<04:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.16     0.4376     0.1236        389        640:  69%|██████▉   | 561/809 [10:59<04:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.16     0.4377     0.1237        281        640:  69%|██████▉   | 562/809 [11:01<04:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.159     0.4378     0.1236        258        640:  70%|██████▉   | 564/809 [11:03<04:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.16     0.4376      0.124        267        640:  70%|██████▉   | 566/809 [11:05<04:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.16     0.4377     0.1239        291        640:  70%|███████   | 567/809 [11:07<04:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.159     0.4379     0.1238        263        640:  70%|███████   | 569/809 [11:09<04:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.158     0.4381     0.1237        135        640:  71%|███████   | 571/809 [11:11<04:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.157     0.4384     0.1236        111        640:  71%|███████   | 573/809 [11:14<04:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.157     0.4385     0.1236        153        640:  71%|███████   | 574/809 [11:15<04:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.156     0.4389     0.1235        238        640:  71%|███████   | 576/809 [11:17<04:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.155     0.4391     0.1233        131        640:  71%|███████▏  | 578/809 [11:19<04:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.155     0.4392     0.1233        201        640:  72%|���██████▏  | 579/809 [11:21<04:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.155     0.4393     0.1236        170        640:  72%|███████▏  | 581/809 [11:23<04:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.154     0.4395     0.1234        199        640:  72%|███████▏  | 583/809 [11:25<04:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.153     0.4399     0.1233        226        640:  72%|███████▏  | 585/809 [11:28<04:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.153     0.4398     0.1234        327        640:  72%|███████▏  | 586/809 [11:29<04:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.152     0.4403     0.1233        125        640:  73%|███████▎  | 588/809 [11:31<04:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.152     0.4404     0.1235        418        640:  73%|███████▎  | 590/809 [11:34<04:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.152     0.4405     0.1234        240        640:  73%|███████▎  | 591/809 [11:35<04:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.153     0.4401     0.1235        519        640:  73%|███████▎  | 593/809 [11:37<04:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.152     0.4402     0.1234        211        640:  74%|███████▎  | 595/809 [11:40<04:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.152     0.4402     0.1234        321        640:  74%|███████▎  | 596/809 [11:41<04:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.152     0.4401     0.1233        326        640:  74%|███████▍  | 598/809 [11:43<04:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.151     0.4403     0.1232        305        640:  74%|███████▍  | 600/809 [11:45<04:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4406      0.123        134        640:  74%|███████▍  | 602/809 [11:48<03:57,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4406     0.1229        262        640:  75%|███████▍  | 603/809 [11:49<03:55,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4406      0.123        121        640:  75%|███████▍  | 605/809 [11:51<03:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4406     0.1229        347        640:  75%|███████▌  | 607/809 [11:54<03:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4404     0.1233        346        640:  75%|███████▌  | 608/809 [11:55<03:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.151     0.4403     0.1236        668        640:  75%|███████▌  | 610/809 [11:57<03:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.151     0.4402     0.1236        102        640:  76%|███████▌  | 612/809 [12:00<04:03,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.151     0.4402     0.1236        357        640:  76%|███████▌  | 613/809 [12:01<03:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.152       0.44     0.1236        363        640:  76%|███████▌  | 615/809 [12:03<03:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.152       0.44      0.124        113        640:  76%|███████▋  | 617/809 [12:06<03:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.152     0.4401     0.1239        334        640:  76%|███████▋  | 618/809 [12:07<03:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4405     0.1238        181        640:  77%|███████▋  | 620/809 [12:09<03:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.151     0.4403     0.1239        460        640:  77%|███████▋  | 622/809 [12:11<03:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.151     0.4403     0.1239        271        640:  77%|███████▋  | 624/809 [12:14<03:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4405     0.1238        126        640:  77%|███████▋  | 625/809 [12:15<03:30,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4407     0.1237         86        640:  78%|███████▊  | 627/809 [12:17<03:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148      0.441     0.1237        157        640:  78%|███████▊  | 629/809 [12:19<03:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4411     0.1237        194        640:  78%|███████▊  | 631/809 [12:22<03:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4411     0.1237        226        640:  78%|███████▊  | 632/809 [12:23<03:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4412     0.1236        246        640:  78%|███████▊  | 634/809 [12:25<03:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4408     0.1241        308        640:  79%|███████▊  | 636/809 [12:28<03:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4408     0.1241        310        640:  79%|███████▊  | 637/809 [12:29<03:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4408     0.1241        248        640:  79%|███████▉  | 639/809 [12:31<03:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149      0.441     0.1241        268        640:  79%|███████▉  | 641/809 [12:34<03:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4411     0.1241        201        640:  79%|███████▉  | 642/809 [12:35<03:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4413      0.124        269        640:  80%|███████▉  | 644/809 [12:37<03:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4412      0.124        341        640:  80%|███████▉  | 646/809 [12:39<03:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4411      0.124        257        640:  80%|████████  | 648/809 [12:42<03:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4412      0.124        414        640:  80%|████████  | 649/809 [12:43<03:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4412      0.124        267        640:  80%|████████  | 651/809 [12:45<03:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4412      0.124        457        640:  81%|████████  | 653/809 [12:48<03:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4412     0.1239        272        640:  81%|████████  | 654/809 [12:49<03:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4414     0.1238        373        640:  81%|████████  | 656/809 [12:51<02:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4415     0.1237        256        640:  81%|████████▏ | 658/809 [12:54<02:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4415     0.1237        242        640:  82%|████████▏ | 660/809 [12:56<02:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4413     0.1238        540        640:  82%|████████▏ | 661/809 [12:57<02:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4414     0.1238        217        640:  82%|████████▏ | 663/809 [12:59<02:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4416     0.1237         84        640:  82%|████████▏ | 665/809 [13:02<02:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4415     0.1237        156        640:  82%|████████▏ | 667/809 [13:04<02:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4416     0.1237        138        640:  83%|████████▎ | 669/809 [13:06<02:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4415     0.1236        249        640:  83%|████████▎ | 670/809 [13:08<02:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4413     0.1237        161        640:  83%|████████▎ | 672/809 [13:10<02:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4411     0.1236        148        640:  83%|████████▎ | 674/809 [13:12<02:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4411     0.1236        264        640:  83%|████████▎ | 675/809 [13:13<02:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4411     0.1236        351        640:  84%|████████▎ | 677/809 [13:16<02:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4413     0.1236        318        640:  84%|████████▍ | 679/809 [13:18<02:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4412     0.1236        420        640:  84%|████████▍ | 681/809 [13:20<02:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4412     0.1236        451        640:  84%|████████▍ | 682/809 [13:22<02:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4412     0.1237        134        640:  85%|████████▍ | 684/809 [13:24<02:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4411     0.1238        571        640:  85%|████████▍ | 686/809 [13:26<02:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4412     0.1238         92        640:  85%|████████▍ | 687/809 [13:28<02:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4411      0.124        367        640:  85%|████████▌ | 689/809 [13:30<02:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4411      0.124        311        640:  85%|████████▌ | 691/809 [13:32<02:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.151     0.4409     0.1242        350        640:  86%|████████▌ | 692/809 [13:34<02:24,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.151     0.4409     0.1242        411        640:  86%|████████▌ | 694/809 [13:36<02:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.151     0.4409     0.1243        174        640:  86%|████████▌ | 696/809 [13:38<02:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.151      0.441     0.1242        227        640:  86%|████████▌ | 697/809 [13:40<02:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.151     0.4411     0.1241        257        640:  86%|████████▋ | 699/809 [13:42<02:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4414     0.1241        133        640:  87%|████████▋ | 701/809 [13:44<02:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4415      0.124        296        640:  87%|████████▋ | 702/809 [13:46<02:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G       1.15     0.4416     0.1241        340        640:  87%|████████▋ | 704/809 [13:48<02:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4418      0.124        140        640:  87%|████████▋ | 706/809 [13:50<01:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4419     0.1239        338        640:  88%|████████▊ | 708/809 [13:52<01:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4421     0.1239        149        640:  88%|████████▊ | 709/809 [13:54<01:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4422     0.1239        305        640:  88%|████████▊ | 711/809 [13:56<01:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4422     0.1238        340        640:  88%|████████▊ | 713/809 [13:58<01:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4423     0.1238        502        640:  88%|████████▊ | 715/809 [14:01<01:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4423     0.1238        245        640:  89%|████████▊ | 716/809 [14:02<01:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4421     0.1238        570        640:  89%|████████▉ | 718/809 [14:04<01:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148      0.442     0.1238        381        640:  89%|████████▉ | 719/809 [14:05<01:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4418      0.124        188        640:  89%|████████▉ | 721/809 [14:08<01:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.149     0.4419      0.124        323        640:  89%|████████▉ | 723/809 [14:10<01:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148     0.4419      0.124        268        640:  90%|████████▉ | 725/809 [14:13<01:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.148      0.442     0.1239        206        640:  90%|████████▉ | 726/809 [14:14<01:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.147     0.4423     0.1238        234        640:  90%|████████▉ | 728/809 [14:16<01:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.147     0.4422     0.1238        249        640:  90%|█████████ | 730/809 [14:19<01:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.147     0.4422     0.1237        206        640:  90%|█████████ | 731/809 [14:20<01:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.147     0.4422     0.1237        359        640:  91%|█████████ | 733/809 [14:22<01:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4425     0.1236        192        640:  91%|█████████ | 735/809 [14:24<01:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4422     0.1238        286        640:  91%|█████████ | 737/809 [14:27<01:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4422     0.1237        179        640:  91%|█████████ | 738/809 [14:28<01:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4422     0.1237        315        640:  91%|█████████▏| 740/809 [14:30<01:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4422     0.1237        344        640:  92%|█████████▏| 742/809 [14:32<01:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4423     0.1237        184        640:  92%|█████████▏| 743/809 [14:34<01:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4422     0.1237        292        640:  92%|█████████▏| 745/809 [14:36<01:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4423     0.1237        142        640:  92%|█████████▏| 747/809 [14:38<01:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4422     0.1236        283        640:  93%|█████████▎| 749/809 [14:41<01:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4422     0.1236        213        640:  93%|█████████▎| 750/809 [14:42<01:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4422     0.1235        190        640:  93%|█████████▎| 752/809 [14:44<01:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4423     0.1235        269        640:  93%|█████████▎| 754/809 [14:47<01:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4423     0.1235        172        640:  93%|█████████▎| 755/809 [14:48<01:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4424     0.1235         84        640:  94%|█████████▎| 757/809 [14:50<01:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4424     0.1235        126        640:  94%|█████████▍| 759/809 [14:52<00:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4423     0.1236        164        640:  94%|█████████▍| 761/809 [14:55<00:55,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4422     0.1236        231        640:  94%|█████████▍| 762/809 [14:56<00:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.146     0.4424     0.1235        211        640:  94%|█████████▍| 764/809 [14:58<00:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.145     0.4426     0.1234        143        640:  95%|█████████▍| 766/809 [15:01<00:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4426     0.1233        179        640:  95%|█████████▍| 767/809 [15:02<00:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.145     0.4425     0.1235        246        640:  95%|█████████▌| 769/809 [15:04<00:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.145     0.4425     0.1234        234        640:  95%|█████████▌| 771/809 [15:06<00:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4426     0.1233         89        640:  96%|█████████▌| 773/809 [15:09<00:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4427     0.1233        307        640:  96%|█████████▌| 774/809 [15:10<00:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4425     0.1233        273        640:  96%|█████████▌| 776/809 [15:12<00:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.145     0.4424     0.1234        366        640:  96%|█████████▌| 778/809 [15:15<00:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4425     0.1233        186        640:  96%|█████████▋| 779/809 [15:16<00:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4426     0.1232        261        640:  97%|█████████▋| 781/809 [15:18<00:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4426     0.1232        247        640:  97%|█████████▋| 783/809 [15:20<00:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.145     0.4424     0.1237        353        640:  97%|█████████▋| 785/809 [15:23<00:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.145     0.4424     0.1237        324        640:  97%|███��█████▋| 786/809 [15:24<00:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4427     0.1236        183        640:  97%|█████████▋| 788/809 [15:26<00:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.144     0.4428     0.1236        379        640:  98%|█████████▊| 790/809 [15:28<00:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.143     0.4429     0.1236        297        640:  98%|█████████▊| 792/809 [15:31<00:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.143     0.4431     0.1235        118        640:  98%|█████████▊| 793/809 [15:32<00:18,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.142     0.4433     0.1235        202        640:  98%|█████████▊| 795/809 [15:34<00:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.141     0.4434     0.1235        340        640:  99%|█████████▊| 797/809 [15:37<00:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.141     0.4433     0.1234        283        640:  99%|█████████▉| 799/809 [15:39<00:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.141     0.4433     0.1234        148        640:  99%|█████████▉| 800/809 [15:40<00:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.141     0.4433     0.1234         78        640:  99%|█████████▉| 802/809 [15:42<00:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.141     0.4432     0.1234        203        640:  99%|█████████▉| 804/809 [15:45<00:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.141     0.4433     0.1233        230        640: 100%|█████████▉| 805/809 [15:46<00:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.141     0.4433     0.1233        128        640: 100%|█████████▉| 807/809 [15:48<00:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     74/300      13.7G      1.142     0.4431     0.1238        566        640: 100%|█████████▉| 808/809 [15:50<00:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  12%|█▏        | 8/69 [00:01<00:14,  4.19it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  23%|██▎       | 16/69 [00:03<00:13,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  35%|███▍      | 24/69 [00:05<00:11,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  46%|████▋     | 32/69 [00:07<00:09,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  58%|█████▊    | 40/69 [00:09<00:07,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  68%|██████▊   | 47/69 [00:11<00:06,  3.50it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  80%|███████▉  | 55/69 [00:13<00:03,  3.87it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  91%|█████████▏| 63/69 [00:15<00:01,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  91%|█████████▏| 63/69 [00:15<00:01,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.6G       1.15     0.4109    0.09489        270        640:   0%|          | 1/809 [00:01<16:15,  1.21s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.321      0.371     0.2002        407        640:   0%|          | 3/809 [00:03<17:12,  1.28s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.136     0.4135      0.164         97        640:   0%|          | 4/809 [00:04<16:24,  1.22s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.142     0.4236     0.1685        238        640:   1%|          | 6/809 [00:07<16:01,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.071     0.4305     0.1456        184        640:   1%|          | 8/809 [00:09<16:19,  1.22s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.06     0.4371     0.1389        246        640:   1%|          | 9/809 [00:10<16:02,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.055     0.4465     0.1315        161        640:   1%|▏         | 11/809 [00:13<16:06,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.091     0.4351     0.1299        328        640:   2%|▏         | 13/809 [00:15<15:46,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.077     0.4391     0.1264        111        640:   2%|▏         | 14/809 [00:16<15:38,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.141     0.4273     0.1545        318        640:   2%|▏         | 16/809 [00:19<15:43,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.155     0.4258     0.1502        207        640:   2%|▏         | 18/809 [00:21<15:20,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.151       0.43     0.1448        295        640:   2%|▏         | 20/809 [00:23<15:31,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.129     0.4359     0.1388        187        640:   3%|▎         | 22/809 [00:26<15:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.125     0.4369     0.1379        239        640:   3%|▎         | 23/809 [00:27<15:42,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.115     0.4407     0.1348        280        640:   3%|▎         | 25/809 [00:29<15:31,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.11     0.4407     0.1325        292        640:   3%|▎         | 27/809 [00:32<15:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.113     0.4385      0.131        327        640:   3%|▎         | 28/809 [00:33<15:09,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4403     0.1295        100        640:   4%|▎         | 30/809 [00:35<15:17,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4425     0.1275        278        640:   4%|▍         | 32/809 [00:38<15:00,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.109     0.4394     0.1285        309        640:   4%|▍         | 34/809 [00:40<14:54,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.113     0.4383     0.1283        383        640:   4%|▍         | 35/809 [00:41<15:06,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.118     0.4366     0.1309        493        640:   5%|▍         | 37/809 [00:43<14:56,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.115     0.4372     0.1293        221        640:   5%|▍         | 39/809 [00:46<14:47,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.116      0.438     0.1281        379        640:   5%|▍         | 40/809 [00:48<14:41,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.116      0.439     0.1274        238        640:   5%|▌         | 42/809 [00:49<14:38,  1.14s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.111     0.4406     0.1264        376        640:   5%|▌         | 44/809 [00:51<14:40,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.115     0.4408     0.1258        342        640:   6%|▌         | 46/809 [00:54<14:36,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.109     0.4428      0.125        259        640:   6%|▌         | 47/809 [00:55<14:32,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4434     0.1241        195        640:   6%|▌         | 49/809 [00:57<14:35,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4433     0.1231        265        640:   6%|▋         | 51/809 [00:59<14:43,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.113     0.4429     0.1227        365        640:   7%|▋         | 53/809 [01:02<14:40,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.114     0.4423     0.1224        408        640:   7%|▋         | 54/809 [01:03<14:28,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.124     0.4393     0.1227        207        640:   7%|▋         | 56/809 [01:05<14:24,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.125     0.4387     0.1219        229        640:   7%|▋         | 58/809 [01:08<14:31,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.121     0.4397      0.121        220        640:   7%|▋         | 60/809 [01:10<14:22,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.116      0.441     0.1204        208        640:   8%|▊         | 61/809 [01:11<14:22,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.111     0.4425     0.1196        253        640:   8%|▊         | 63/809 [01:13<14:25,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.109     0.4435     0.1204        466        640:   8%|▊         | 65/809 [01:16<14:22,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4446     0.1197        140        640:   8%|▊         | 66/809 [01:17<14:13,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4447      0.119        280        640:   8%|▊         | 68/809 [01:19<14:18,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.098     0.4467     0.1183        167        640:   9%|▊         | 70/809 [01:21<14:09,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4459     0.1186        139        640:   9%|▉         | 72/809 [01:24<14:07,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4458     0.1184        292        640:   9%|▉         | 73/809 [01:25<14:13,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.09     0.4488     0.1175        234        640:   9%|▉         | 75/809 [01:27<14:32,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4471      0.118        384        640:  10%|▉         | 77/809 [01:30<14:23,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4473     0.1175        285        640:  10%|▉         | 78/809 [01:31<14:23,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102      0.446     0.1198        188        640:  10%|▉         | 80/809 [01:33<14:18,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107      0.445     0.1202        132        640:  10%|█         | 82/809 [01:36<14:07,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4455     0.1197        189        640:  10%|█         | 84/809 [01:38<14:28,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4468     0.1194        235        640:  11%|█         | 85/809 [01:39<14:31,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.093     0.4491     0.1182        135        640:  11%|█         | 87/809 [01:41<14:08,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4488     0.1183        130        640:  11%|█         | 89/809 [01:44<13:59,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.098      0.448     0.1201        466        640:  11%|█         | 90/809 [01:45<13:56,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4477     0.1202        241        640:  11%|█▏        | 92/809 [01:47<13:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4495     0.1192        369        640:  12%|█▏        | 94/809 [01:50<13:53,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4507     0.1182        149        640:  12%|█▏        | 96/809 [01:52<13:46,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.085     0.4515     0.1178        216        640:  12%|█▏        | 97/809 [01:53<13:50,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.081     0.4521     0.1171        306        640:  12%|█▏        | 99/809 [01:56<13:53,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.072     0.4543     0.1165        104        640:  12%|█▏        | 101/809 [01:58<13:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.069     0.4555     0.1162        135        640:  13%|█▎        | 102/809 [01:59<13:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.064     0.4567     0.1152        145        640:  13%|█▎        | 104/809 [02:01<13:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.063     0.4568     0.1149        417        640:  13%|█▎        | 106/809 [02:04<13:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.063     0.4575     0.1151        209        640:  13%|█▎        | 108/809 [02:06<13:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.065     0.4569     0.1156        394        640:  13%|█▎        | 109/809 [02:07<13:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.07     0.4558     0.1161        450        640:  14%|█▎        | 111/809 [02:10<13:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.069     0.4565     0.1157        210        640:  14%|█▍        | 113/809 [02:12<13:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.069     0.4568     0.1155        191        640:  14%|█▍        | 114/809 [02:13<13:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.073     0.4564     0.1171        307        640:  14%|█▍        | 116/809 [02:16<13:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.068     0.4579     0.1164        259        640:  15%|█▍        | 118/809 [02:18<13:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.065     0.4586      0.116        167        640:  15%|█▍        | 119/809 [02:19<13:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.064     0.4592     0.1155        277        640:  15%|█▍        | 121/809 [02:21<13:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.066     0.4587     0.1157        142        640:  15%|█▌        | 123/809 [02:24<13:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.063     0.4593     0.1149        198        640:  15%|█▌        | 125/809 [02:26<13:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.063     0.4594     0.1146        217        640:  16%|█▌        | 126/809 [02:27<13:08,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.063     0.4598     0.1144        142        640:  16%|█▌        | 128/809 [02:30<13:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.061     0.4605     0.1141        223        640:  16%|█▌        | 130/809 [02:32<13:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.059     0.4608     0.1139        160        640:  16%|█▌        | 131/809 [02:33<13:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.056     0.4615     0.1133        138        640:  16%|█▋        | 133/809 [02:35<13:00,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.053     0.4624     0.1127        207        640:  17%|█▋        | 135/809 [02:38<12:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.055     0.4623     0.1137        196        640:  17%|█▋        | 137/809 [02:40<12:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.056     0.4626     0.1138        298        640:  17%|█▋        | 138/809 [02:41<12:50,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.054     0.4632     0.1137        274        640:  17%|█▋        | 140/809 [02:43<12:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.055     0.4631     0.1134        211        640:  18%|█▊        | 142/809 [02:46<13:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.054     0.4637     0.1131        237        640:  18%|█▊        | 144/809 [02:48<12:45,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.058     0.4627     0.1153        630        640:  18%|█▊        | 145/809 [02:49<13:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.056     0.4636     0.1149        245        640:  18%|█▊        | 147/809 [02:52<12:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.062     0.4619     0.1155        470        640:  18%|█▊        | 149/809 [02:54<12:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.063     0.4613     0.1155        217        640:  19%|█▊        | 150/809 [02:55<12:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.068     0.4601     0.1157        292        640:  19%|█▉        | 152/809 [02:57<12:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.076     0.4579     0.1171        255        640:  19%|█▉        | 154/809 [03:00<12:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.079     0.4572     0.1172        341        640:  19%|█▉        | 156/809 [03:02<12:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.077     0.4578      0.117         92        640:  19%|█▉        | 157/809 [03:03<12:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.078     0.4575     0.1169        371        640:  20%|█▉        | 159/809 [03:06<13:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.078     0.4574     0.1167        202        640:  20%|█▉        | 161/809 [03:08<12:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.081     0.4566     0.1173        499        640:  20%|██        | 162/809 [03:09<12:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.082     0.4563     0.1172        320        640:  20%|██        | 164/809 [03:12<12:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.086     0.4551     0.1174        353        640:  21%|██        | 166/809 [03:14<12:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.087      0.455     0.1174        235        640:  21%|██        | 168/809 [03:16<12:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4546     0.1173        335        640:  21%|██        | 169/809 [03:17<12:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.091     0.4537     0.1174        392        640:  21%|██        | 171/809 [03:20<12:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.09     0.4541     0.1172        232        640:  21%|██▏       | 173/809 [03:22<12:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.09     0.4547     0.1173        437        640:  22%|██▏       | 175/809 [03:25<12:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.092     0.4544     0.1183        377        640:  22%|██▏       | 176/809 [03:26<12:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.095     0.4536     0.1185        339        640:  22%|██▏       | 178/809 [03:28<12:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094      0.454     0.1181        341        640:  22%|██▏       | 180/809 [03:30<12:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.092     0.4546     0.1179        240        640:  22%|██▏       | 182/809 [03:33<12:02,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.093     0.4544     0.1183        359        640:  23%|██▎       | 183/809 [03:34<12:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.095     0.4542     0.1183        401        640:  23%|██▎       | 185/809 [03:36<12:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4548     0.1181        223        640:  23%|██▎       | 187/809 [03:39<12:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.095     0.4549     0.1184        254        640:  23%|██▎       | 189/809 [03:41<11:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4555     0.1181        170        640:  23%|██▎       | 190/809 [03:42<11:52,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4561     0.1181        277        640:  24%|██▎       | 192/809 [03:44<11:49,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4553     0.1193        443        640:  24%|██▍       | 194/809 [03:47<12:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4548     0.1196        331        640:  24%|██▍       | 195/809 [03:48<12:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4545     0.1195        340        640:  24%|██▍       | 197/809 [03:50<11:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4537       0.12        393        640:  25%|██▍       | 199/809 [03:52<11:40,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107      0.454       0.12        268        640:  25%|██▍       | 201/809 [03:55<11:34,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.109     0.4538     0.1201        293        640:  25%|██▍       | 202/809 [03:56<11:39,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.109     0.4535     0.1198        215        640:  25%|██▌       | 204/809 [03:58<11:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4542     0.1194        157        640:  25%|██▌       | 206/809 [04:01<11:31,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108      0.454     0.1207        530        640:  26%|██▌       | 207/809 [04:02<11:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.11     0.4536     0.1207        315        640:  26%|██▌       | 209/809 [04:04<11:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.109     0.4541     0.1204        375        640:  26%|██▌       | 211/809 [04:06<11:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.109     0.4541     0.1205        200        640:  26%|██▋       | 213/809 [04:09<11:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.109     0.4541     0.1204        203        640:  26%|██▋       | 214/809 [04:10<11:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4547     0.1201        114        640:  27%|██▋       | 216/809 [04:12<11:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.109     0.4541     0.1198        252        640:  27%|██▋       | 218/809 [04:15<11:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.111     0.4536     0.1198        400        640:  27%|██▋       | 219/809 [04:16<11:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4545     0.1196        274        640:  27%|██▋       | 221/809 [04:18<11:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4548     0.1193        161        640:  28%|██▊       | 223/809 [04:21<11:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103      0.456     0.1189        179        640:  28%|██▊       | 225/809 [04:23<11:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4565     0.1187         88        640:  28%|██▊       | 226/809 [04:24<11:10,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4563     0.1193        327        640:  28%|██▊       | 228/809 [04:26<11:10,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101      0.457     0.1191        386        640:  28%|██▊       | 230/809 [04:29<11:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4577     0.1186        170        640:  29%|██▊       | 232/809 [04:31<11:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.098     0.4577     0.1185        277        640:  29%|██▉       | 233/809 [04:32<11:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.098     0.4577     0.1188        186        640:  29%|██▉       | 235/809 [04:34<11:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4573     0.1188        262        640:  29%|██▉       | 237/809 [04:37<11:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4573     0.1189        308        640:  29%|██▉       | 238/809 [04:38<11:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4566     0.1195        316        640:  30%|██▉       | 240/809 [04:40<11:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4561     0.1196        439        640:  30%|██▉       | 242/809 [04:43<11:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4561     0.1193        267        640:  30%|███       | 244/809 [04:45<10:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4558     0.1194        472        640:  30%|███       | 245/809 [04:46<10:50,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4557     0.1193        223        640:  31%|███       | 247/809 [04:49<11:28,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106      0.456     0.1191        138        640:  31%|███       | 249/809 [04:51<11:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4562     0.1189        338        640:  31%|███       | 250/809 [04:52<11:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4554     0.1195        415        640:  31%|███       | 252/809 [04:55<11:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4556     0.1191        142        640:  31%|███▏      | 254/809 [04:57<10:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4552      0.119        296        640:  32%|███▏      | 255/809 [04:58<10:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.11     0.4545      0.119        197        640:  32%|███▏      | 257/809 [05:00<10:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.109     0.4545     0.1187        175        640:  32%|███▏      | 259/809 [05:03<11:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4547     0.1185        115        640:  32%|███▏      | 260/809 [05:04<10:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4549     0.1182        235        640:  32%|███▏      | 262/809 [05:06<10:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4549     0.1185        179        640:  33%|███▎      | 264/809 [05:09<10:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4548     0.1183        242        640:  33%|███▎      | 266/809 [05:11<10:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4551     0.1182        220        640:  33%|███▎      | 267/809 [05:12<10:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4551     0.1181        211        640:  33%|███▎      | 269/809 [05:15<10:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4553     0.1179        381        640:  33%|███▎      | 271/809 [05:17<10:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4555     0.1179        170        640:  34%|███▎      | 272/809 [05:18<10:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4554     0.1178        191        640:  34%|███▍      | 274/809 [05:20<10:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4555     0.1177        116        640:  34%|███▍      | 276/809 [05:23<10:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4554     0.1177        269        640:  34%|███▍      | 277/809 [05:24<10:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.111     0.4548     0.1186        498        640:  34%|███▍      | 279/809 [05:26<10:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.112     0.4547     0.1185        455        640:  35%|███▍      | 281/809 [05:29<10:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.112     0.4547     0.1188        294        640:  35%|███▍      | 283/809 [05:31<10:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.112     0.4545     0.1187        323        640:  35%|███▌      | 284/809 [05:32<10:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.113     0.4544     0.1186        193        640:  35%|███▌      | 286/809 [05:35<10:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.116     0.4534     0.1194        224        640:  36%|███▌      | 288/809 [05:37<10:00,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.119     0.4529     0.1194        122        640:  36%|███▌      | 290/809 [05:39<10:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.118      0.453     0.1193        202        640:  36%|███▌      | 291/809 [05:40<10:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.119     0.4528     0.1191        238        640:  36%|███▌      | 293/809 [05:43<10:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.118     0.4531     0.1189        187        640:  36%|███▋      | 295/809 [05:45<09:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.118     0.4532     0.1188        252        640:  37%|███▋      | 296/809 [05:46<09:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.116      0.454     0.1187        104        640:  37%|███▋      | 298/809 [05:49<10:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.116     0.4542     0.1186        128        640:  37%|███▋      | 300/809 [05:51<10:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.117     0.4541     0.1187        414        640:  37%|███▋      | 301/809 [05:52<10:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.115     0.4549     0.1184        113        640:  37%|███▋      | 303/809 [05:55<09:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.116     0.4547     0.1186        462        640:  38%|███▊      | 305/809 [05:57<09:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.114     0.4556     0.1183        203        640:  38%|███▊      | 307/809 [05:59<09:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.114     0.4558     0.1181        253        640:  38%|███▊      | 308/809 [06:00<09:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.113     0.4557     0.1181        245        640:  38%|███▊      | 310/809 [06:03<09:34,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.113     0.4558     0.1181        223        640:  39%|███▊      | 312/809 [06:05<09:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.114     0.4556     0.1182        329        640:  39%|███▉      | 314/809 [06:07<09:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.112     0.4558      0.118        228        640:  39%|███▉      | 316/809 [06:10<09:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.111     0.4561     0.1178        309        640:  39%|███▉      | 317/809 [06:11<09:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4567     0.1175        242        640:  39%|███▉      | 319/809 [06:13<09:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4571     0.1172        210        640:  40%|███▉      | 321/809 [06:16<09:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4574     0.1171        136        640:  40%|███▉      | 322/809 [06:17<09:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4575     0.1169        109        640:  40%|████      | 324/809 [06:19<09:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4576     0.1167        317        640:  40%|████      | 326/809 [06:21<09:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4582     0.1166        213        640:  41%|████      | 328/809 [06:24<09:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4585     0.1164        189        640:  41%|████      | 329/809 [06:25<09:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4583     0.1163        156        640:  41%|████      | 331/809 [06:27<09:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4585      0.116         89        640:  41%|████      | 333/809 [06:30<09:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4587     0.1159        240        640:  41%|████▏     | 334/809 [06:31<09:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4588     0.1157        211        640:  42%|████▏     | 336/809 [06:33<09:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.098     0.4591     0.1155         76        640:  42%|████▏     | 338/809 [06:35<09:00,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.097     0.4593     0.1153        271        640:  42%|████▏     | 340/809 [06:38<08:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4594     0.1152        180        640:  42%|████▏     | 341/809 [06:39<08:55,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.095     0.4596      0.115        163        640:  42%|████▏     | 343/809 [06:41<08:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.095     0.4597     0.1147        286        640:  43%|████▎     | 345/809 [06:44<08:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4595     0.1145        313        640:  43%|████▎     | 347/809 [06:46<09:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4597     0.1144        151        640:  43%|████▎     | 348/809 [06:47<09:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.093     0.4601     0.1144        113        640:  43%|████▎     | 350/809 [06:49<08:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.092     0.4603     0.1143        163        640:  44%|████▎     | 352/809 [06:52<08:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.091     0.4607     0.1141        140        640:  44%|████▎     | 353/809 [06:53<08:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.091     0.4607      0.114        311        640:  44%|████▍     | 355/809 [06:55<09:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.092     0.4604      0.114        365        640:  44%|████▍     | 357/809 [06:58<09:18,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.093     0.4602      0.114        428        640:  44%|████▍     | 358/809 [06:59<09:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.092     0.4603     0.1139        312        640:  44%|████▍     | 360/809 [07:02<08:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4598     0.1141        495        640:  45%|████▍     | 362/809 [07:04<08:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.095     0.4596     0.1147        479        640:  45%|████▍     | 363/809 [07:05<08:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.095     0.4592     0.1147        206        640:  45%|████▌     | 365/809 [07:07<08:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.098     0.4586     0.1152        401        640:  45%|████▌     | 367/809 [07:10<08:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4583     0.1152        150        640:  46%|████▌     | 369/809 [07:12<08:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.098     0.4583     0.1152        160        640:  46%|████▌     | 370/809 [07:13<08:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.097     0.4585      0.115        306        640:  46%|████▌     | 372/809 [07:15<08:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.098     0.4581     0.1151        310        640:  46%|████▌     | 374/809 [07:18<08:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.097     0.4583      0.115        208        640:  46%|████▋     | 375/809 [07:19<08:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4586     0.1148        139        640:  47%|████▋     | 377/809 [07:21<08:18,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4586     0.1146        219        640:  47%|████▋     | 379/809 [07:24<08:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4584     0.1145        470        640:  47%|████▋     | 381/809 [07:26<08:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4585     0.1144        216        640:  47%|████▋     | 382/809 [07:27<08:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4585     0.1143        141        640:  47%|████▋     | 384/809 [07:29<08:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4586     0.1143        426        640:  48%|████▊     | 386/809 [07:32<08:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4584     0.1143        401        640:  48%|████▊     | 387/809 [07:33<08:37,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4584     0.1143        224        640:  48%|████▊     | 389/809 [07:36<08:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.097      0.458     0.1149        245        640:  48%|████▊     | 391/809 [07:38<08:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4576     0.1153        511        640:  48%|████▊     | 392/809 [07:39<08:31,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4576     0.1151        140        640:  49%|████▊     | 394/809 [07:42<08:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.098     0.4578     0.1151        191        640:  49%|████▉     | 396/809 [07:44<08:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.098     0.4575     0.1151        351        640:  49%|████▉     | 397/809 [07:45<08:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.097     0.4577     0.1151        539        640:  49%|████▉     | 399/809 [07:48<08:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4582     0.1149        205        640:  50%|████▉     | 401/809 [07:50<07:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4581     0.1148        341        640:  50%|████▉     | 402/809 [07:51<07:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4581     0.1148        256        640:  50%|████▉     | 404/809 [07:54<08:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096      0.458     0.1148        169        640:  50%|█████     | 406/809 [07:56<07:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4581     0.1147         94        640:  50%|█████     | 407/809 [07:57<07:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4584     0.1145        119        640:  51%|█████     | 409/809 [07:59<07:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.093     0.4586     0.1143        188        640:  51%|█████     | 411/809 [08:02<08:10,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.093     0.4587     0.1142        133        640:  51%|█████     | 412/809 [08:03<07:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4584     0.1143        216        640:  51%|█████     | 414/809 [08:06<07:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.093     0.4585     0.1142        210        640:  51%|█████▏    | 416/809 [08:08<07:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4583     0.1142        359        640:  52%|█████▏    | 418/809 [08:10<07:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4583     0.1141        205        640:  52%|█████▏    | 419/809 [08:11<07:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.092     0.4588     0.1139        140        640:  52%|█████▏    | 421/809 [08:14<07:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.091     0.4588     0.1138        104        640:  52%|█████▏    | 423/809 [08:16<07:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.091     0.4586     0.1137        259        640:  52%|█████▏    | 424/809 [08:17<07:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.09     0.4589     0.1137        279        640:  53%|█████▎    | 426/809 [08:19<07:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.09     0.4588     0.1136        313        640:  53%|█████▎    | 428/809 [08:22<07:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.09     0.4588     0.1135        369        640:  53%|█████▎    | 430/809 [08:24<07:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.09     0.4588     0.1134        279        640:  53%|█████▎    | 431/809 [08:25<07:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.089     0.4595     0.1132         71        640:  54%|█████▎    | 433/809 [08:28<07:12,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.089     0.4593     0.1132        321        640:  54%|█████▍    | 435/809 [08:30<07:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.089     0.4595     0.1131         55        640:  54%|█████▍    | 436/809 [08:31<07:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.089     0.4596     0.1131        125        640:  54%|█████▍    | 438/809 [08:33<07:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.089     0.4597     0.1131        120        640:  54%|█████▍    | 440/809 [08:36<07:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4599      0.113        186        640:  55%|█████▍    | 442/809 [08:38<07:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.089     0.4599      0.113        294        640:  55%|█████▍    | 443/809 [08:39<07:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4602     0.1129        229        640:  55%|█████▌    | 445/809 [08:42<07:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.089       0.46     0.1131        342        640:  55%|█████▌    | 447/809 [08:44<06:57,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4605      0.113        194        640:  56%|█████▌    | 449/809 [08:46<07:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4606     0.1129        248        640:  56%|█████▌    | 450/809 [08:47<07:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4604     0.1129        420        640:  56%|█████▌    | 452/809 [08:50<07:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4604     0.1129        165        640:  56%|█████▌    | 454/809 [08:52<06:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4605     0.1128        187        640:  56%|█████▌    | 455/809 [08:53<06:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4604     0.1128        272        640:  56%|█████▋    | 457/809 [08:56<06:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4603     0.1127        182        640:  57%|█████▋    | 459/809 [08:58<06:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.087     0.4605     0.1126        134        640:  57%|█████▋    | 460/809 [08:59<07:12,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.089     0.4602     0.1128        200        640:  57%|█████▋    | 462/809 [09:02<06:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.087     0.4606     0.1126        156        640:  57%|█████▋    | 464/809 [09:04<06:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.087      0.461     0.1125        249        640:  58%|█████▊    | 466/809 [09:07<07:02,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.087     0.4608     0.1125        299        640:  58%|█████▊    | 467/809 [09:08<06:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.086     0.4612     0.1124        119        640:  58%|█████▊    | 469/809 [09:10<06:54,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4611     0.1136        587        640:  58%|█████▊    | 471/809 [09:13<06:51,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4612     0.1136        272        640:  58%|█████▊    | 472/809 [09:14<06:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.088     0.4611     0.1138        328        640:  59%|█████▊    | 474/809 [09:16<06:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.09     0.4608     0.1147        386        640:  59%|█████▉    | 476/809 [09:19<06:52,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G       1.09     0.4607     0.1147        194        640:  59%|█████▉    | 477/809 [09:20<06:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.092     0.4604     0.1157        837        640:  59%|█████▉    | 479/809 [09:22<06:56,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.092     0.4606     0.1156        383        640:  59%|█████▉    | 481/809 [09:25<06:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.092     0.4605     0.1155        231        640:  60%|█████▉    | 482/809 [09:26<06:43,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.094     0.4598     0.1177        557        640:  60%|█████▉    | 484/809 [09:29<06:49,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.095     0.4597     0.1178        200        640:  60%|█████▉    | 485/809 [09:30<06:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.097     0.4591     0.1189        831        640:  60%|██████    | 487/809 [09:32<06:56,  1.29s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4595     0.1187        135        640:  60%|██████    | 489/809 [09:35<06:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4595     0.1187        256        640:  61%|██████    | 490/809 [09:36<06:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4595     0.1187        238        640:  61%|██████    | 492/809 [09:38<06:32,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4597     0.1186         88        640:  61%|██████    | 494/809 [09:41<06:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.097     0.4595     0.1187        392        640:  61%|██████    | 495/809 [09:42<06:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.097     0.4594     0.1187        250        640:  61%|██████▏   | 497/809 [09:44<06:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4595     0.1187        175        640:  62%|██████▏   | 499/809 [09:47<06:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.096     0.4595     0.1186        202        640:  62%|██████▏   | 501/809 [09:49<05:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.098     0.4592     0.1189        360        640:  62%|██████▏   | 502/809 [09:50<05:52,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4589     0.1189        458        640:  62%|██████▏   | 504/809 [09:52<05:48,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4585      0.119        290        640:  63%|██████▎   | 506/809 [09:55<05:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4583     0.1191        329        640:  63%|██████▎   | 507/809 [09:56<06:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102      0.458     0.1193        276        640:  63%|██████▎   | 509/809 [09:58<05:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4582     0.1192        161        640:  63%|██████▎   | 511/809 [10:01<05:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4585     0.1191        260        640:  63%|██████▎   | 513/809 [10:03<05:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4582     0.1194        331        640:  64%|██████▎   | 514/809 [10:04<05:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4586     0.1192        175        640:  64%|██████▍   | 516/809 [10:06<05:37,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4584     0.1192        271        640:  64%|██████▍   | 518/809 [10:09<05:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4584     0.1191        218        640:  64%|██████▍   | 520/809 [10:11<05:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4585     0.1191        307        640:  64%|██████���   | 521/809 [10:12<05:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4589      0.119        150        640:  65%|██████▍   | 523/809 [10:15<05:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4586     0.1194        414        640:  65%|██████▍   | 525/809 [10:17<05:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4587     0.1194        210        640:  65%|██████▌   | 526/809 [10:18<05:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4586     0.1194        242        640:  65%|██████▌   | 528/809 [10:20<05:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4586     0.1194        446        640:  66%|██████▌   | 530/809 [10:23<05:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4585     0.1193        235        640:  66%|██████▌   | 532/809 [10:25<05:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4585     0.1192        203        640:  66%|██████▌   | 533/809 [10:26<05:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4586     0.1192        164        640:  66%|██████▌   | 535/809 [10:28<05:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4588     0.1191         78        640:  66%|██████▋   | 537/809 [10:31<05:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4586     0.1191        356        640:  67%|██████▋   | 538/809 [10:32<05:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4585     0.1194        402        640:  67%|██████▋   | 540/809 [10:34<05:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4586     0.1193        189        640:  67%|██████▋   | 542/809 [10:37<05:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4585     0.1193        252        640:  67%|██████▋   | 544/809 [10:39<05:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4584     0.1193        444        640:  67%|██████▋   | 545/809 [10:40<05:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4584     0.1192        266        640:  68%|██████▊   | 547/809 [10:42<05:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4585     0.1191        223        640:  68%|██████▊   | 549/809 [10:45<05:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4585      0.119        256        640:  68%|██████▊   | 551/809 [10:47<04:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.099     0.4586     0.1189        331        640:  68%|██████▊   | 552/809 [10:48<04:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1     0.4582     0.1192        308        640:  68%|██████▊   | 554/809 [10:51<04:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G        1.1      0.458     0.1193        330        640:  69%|██████▊   | 556/809 [10:53<04:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101      0.458     0.1194        391        640:  69%|██████▉   | 557/809 [10:54<04:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4579     0.1194        368        640:  69%|██████▉   | 559/809 [10:56<04:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4577     0.1194        448        640:  69%|██████▉   | 561/809 [10:59<04:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4574     0.1195        453        640:  70%|██████▉   | 563/809 [11:01<04:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4574     0.1195        269        640:  70%|██████▉   | 564/809 [11:02<04:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4576     0.1193        217        640:  70%|██████▉   | 566/809 [11:05<04:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4574     0.1193        428        640:  70%|███████   | 568/809 [11:07<04:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4574     0.1193        158        640:  70%|███████   | 569/809 [11:08<04:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4572     0.1193        222        640:  71%|███████   | 571/809 [11:10<04:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103      0.457     0.1195        218        640:  71%|███████   | 573/809 [11:13<04:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103      0.457     0.1195        253        640:  71%|███████   | 574/809 [11:14<04:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103      0.457     0.1194        191        640:  71%|███████   | 576/809 [11:16<04:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4569     0.1195        221        640:  71%|███████▏  | 578/809 [11:19<04:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4567     0.1198        151        640:  72%|███████▏  | 580/809 [11:21<04:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4565     0.1198        383        640:  72%|███████▏  | 581/809 [11:22<04:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4565     0.1199        270        640:  72%|███████▏  | 583/809 [11:25<04:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4566     0.1197         53        640:  72%|███████▏  | 585/809 [11:27<04:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4567     0.1197        287        640:  72%|███████▏  | 586/809 [11:28<04:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4567     0.1196         93        640:  73%|███████▎  | 588/809 [11:31<04:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4569     0.1196        272        640:  73%|███████▎  | 590/809 [11:33<04:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4571     0.1196        230        640:  73%|███████▎  | 592/809 [11:35<04:09,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4572     0.1195        272        640:  73%|███████▎  | 593/809 [11:36<04:09,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102      0.457     0.1195        326        640:  74%|███████▎  | 595/809 [11:39<04:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4572     0.1195        272        640:  74%|███████▍  | 597/809 [11:41<04:04,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.101     0.4573     0.1194        137        640:  74%|███████▍  | 598/809 [11:42<04:02,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4572     0.1194        296        640:  74%|███████▍  | 600/809 [11:45<04:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4572     0.1193        268        640:  74%|███████▍  | 602/809 [11:47<04:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4569     0.1199        527        640:  75%|███████▍  | 603/809 [11:48<04:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4567     0.1199        347        640:  75%|███████▍  | 605/809 [11:50<04:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4565     0.1198        302        640:  75%|███████▌  | 607/809 [11:53<03:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4563     0.1198        437        640:  75%|███████▌  | 609/809 [11:55<03:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4563     0.1198        175        640:  75%|███████▌  | 610/809 [11:56<03:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4566     0.1197        132        640:  76%|███████▌  | 612/809 [11:59<03:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4564     0.1198        277        640:  76%|███████▌  | 614/809 [12:01<03:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4564     0.1198        177        640:  76%|███████▌  | 616/809 [12:03<03:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4564     0.1198        349        640:  76%|███████▋  | 618/809 [12:06<03:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4567     0.1197        103        640:  77%|███████▋  | 619/809 [12:07<03:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4565     0.1197        315        640:  77%|███████▋  | 621/809 [12:09<03:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4564     0.1197        217        640:  77%|███████▋  | 623/809 [12:12<03:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4564     0.1197        138        640:  77%|███████▋  | 625/809 [12:14<03:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4564     0.1197        269        640:  77%|███████▋  | 626/809 [12:15<03:31,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4563     0.1197        187        640:  78%|███████▊  | 628/809 [12:17<03:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106      0.456     0.1199        313        640:  78%|███████▊  | 630/809 [12:20<03:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4559     0.1199        284        640:  78%|███████▊  | 631/809 [12:21<03:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4558     0.1199        296        640:  78%|███████▊  | 633/809 [12:23<03:22,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4556     0.1199        486        640:  78%|███████▊  | 635/809 [12:26<03:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.108     0.4556     0.1199        343        640:  79%|███████▊  | 637/809 [12:28<03:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4557     0.1198        186        640:  79%|███████▉  | 638/809 [12:29<03:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4558     0.1197        294        640:  79%|███████▉  | 640/809 [12:31<03:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4561     0.1196        133        640:  79%|███████▉  | 642/809 [12:34<03:12,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4563     0.1195        117        640:  79%|███████▉  | 643/809 [12:35<03:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106      0.456     0.1196        259        640:  80%|███████▉  | 645/809 [12:37<03:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4562     0.1195        145        640:  80%|███████▉  | 647/809 [12:39<03:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106      0.456     0.1196        416        640:  80%|████████  | 649/809 [12:42<03:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4562     0.1195        172        640:  80%|████████  | 650/809 [12:43<03:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4561     0.1195        400        640:  81%|████████  | 652/809 [12:45<03:00,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105      0.456     0.1194        305        640:  81%|████████  | 654/809 [12:48<03:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105      0.456     0.1194        114        640:  81%|████████  | 656/809 [12:50<02:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4559     0.1195        518        640:  81%|████████  | 657/809 [12:51<02:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4559     0.1194        213        640:  81%|████████▏ | 659/809 [12:53<02:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105      0.456     0.1194        233        640:  82%|████████▏ | 661/809 [12:56<02:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105      0.456     0.1193        228        640:  82%|████████▏ | 662/809 [12:57<02:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4559     0.1194        294        640:  82%|████████▏ | 664/809 [12:59<02:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4558     0.1199        150        640:  82%|████████▏ | 666/809 [13:02<02:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4558     0.1199        234        640:  82%|████████▏ | 667/809 [13:03<02:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4558     0.1197        176        640:  83%|████████▎ | 669/809 [13:05<02:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4559     0.1196        303        640:  83%|████████▎ | 671/809 [13:08<02:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4559     0.1196        163        640:  83%|████████▎ | 673/809 [13:10<02:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4556     0.1197        351        640:  83%|████████▎ | 674/809 [13:11<02:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4554     0.1198        466        640:  84%|████████▎ | 676/809 [13:14<02:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4555     0.1197         97        640:  84%|████████▍ | 678/809 [13:16<02:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4554     0.1197        372        640:  84%|████████▍ | 679/809 [13:17<02:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4556     0.1196        187        640:  84%|████████▍ | 681/809 [13:19<02:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4557     0.1195        239        640:  84%|████████▍ | 683/809 [13:22<02:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4558     0.1195        232        640:  85%|████████▍ | 684/809 [13:23<02:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4558     0.1196        395        640:  85%|████████▍ | 686/809 [13:25<02:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4559     0.1195        336        640:  85%|████████▌ | 688/809 [13:28<02:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104      0.456     0.1194        246        640:  85%|████████▌ | 690/809 [13:30<02:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104      0.456     0.1196        292        640:  85%|████████▌ | 691/809 [13:31<02:24,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104      0.456     0.1196        233        640:  86%|████████▌ | 693/809 [13:34<02:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4559     0.1195        174        640:  86%|████████▌ | 695/809 [13:36<02:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4559     0.1195        282        640:  86%|████████▌ | 696/809 [13:37<02:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4559     0.1195        308        640:  86%|████████▋ | 698/809 [13:40<02:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4558     0.1196        300        640:  87%|████████▋ | 700/809 [13:42<02:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104      0.456     0.1195        234        640:  87%|████████▋ | 701/809 [13:43<02:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4559     0.1197        349        640:  87%|████████▋ | 703/809 [13:46<02:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4559     0.1196        157        640:  87%|████████▋ | 705/809 [13:48<02:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103      0.456     0.1195         65        640:  87%|████████▋ | 706/809 [13:49<02:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4559     0.1197        276        640:  88%|████████▊ | 708/809 [13:52<02:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4558     0.1197        208        640:  88%|████████▊ | 710/809 [13:54<01:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4557     0.1197        311        640:  88%|████████▊ | 711/809 [13:55<01:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4554     0.1202        299        640:  88%|████████▊ | 713/809 [13:58<01:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4555     0.1201        347        640:  88%|████████▊ | 715/809 [14:00<01:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4553     0.1202        353        640:  89%|████████▊ | 716/809 [14:01<01:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106      0.455     0.1205        387        640:  89%|████████▉ | 718/809 [14:03<01:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4552     0.1204        157        640:  89%|████████▉ | 720/809 [14:06<01:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4552     0.1203        327        640:  89%|████████▉ | 722/809 [14:08<01:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4553     0.1203         93        640:  89%|████████▉ | 723/809 [14:09<01:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4551     0.1208        162        640:  90%|████████▉ | 725/809 [14:12<01:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4551     0.1208        234        640:  90%|████████▉ | 727/809 [14:14<01:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4553     0.1207        100        640:  90%|█████████ | 729/809 [14:16<01:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4552     0.1207        281        640:  90%|█████████ | 730/809 [14:17<01:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4553     0.1206        363        640:  90%|█████████ | 732/809 [14:20<01:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4555     0.1206        157        640:  91%|█████████ | 734/809 [14:22<01:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4555     0.1206        262        640:  91%|█████████ | 735/809 [14:23<01:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4555     0.1206        168        640:  91%|█████████ | 737/809 [14:26<01:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4555     0.1205        213        640:  91%|█████████▏| 739/809 [14:28<01:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4554     0.1205        280        640:  92%|█████████▏| 741/809 [14:30<01:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4553     0.1205        301        640:  92%|█████████▏| 742/809 [14:31<01:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4553     0.1204        282        640:  92%|█████████▏| 744/809 [14:34<01:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4555     0.1203        145        640:  92%|█████████▏| 746/809 [14:36<01:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4556     0.1202        173        640:  92%|█████████▏| 747/809 [14:37<01:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4556     0.1202        294        640:  93%|█████████▎| 749/809 [14:40<01:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4555     0.1202        236        640:  93%|█████████▎| 751/809 [14:42<01:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4556     0.1201        438        640:  93%|█████████▎| 752/809 [14:43<01:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4555     0.1202        491        640:  93%|█████████▎| 754/809 [14:45<01:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4557     0.1201        232        640:  93%|█████████▎| 756/809 [14:48<01:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4559     0.1202        307        640:  94%|█████████▎| 758/809 [14:50<00:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4559     0.1201        300        640:  94%|█████████▍| 759/809 [14:51<00:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4559     0.1201        175        640:  94%|█████████▍| 761/809 [14:54<00:55,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.102     0.4558     0.1202        340        640:  94%|█████████▍| 763/809 [14:56<00:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4556     0.1203        444        640:  95%|█████████▍| 765/809 [14:58<00:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4556     0.1203        258        640:  95%|█████████▍| 767/809 [15:01<00:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4555     0.1203        212        640:  95%|█████████▍| 768/809 [15:02<00:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.103     0.4554     0.1203        324        640:  95%|█████████▌| 770/809 [15:04<00:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4551     0.1206        344        640:  95%|█████████▌| 772/809 [15:07<00:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4552     0.1206        143        640:  96%|█████████▌| 773/809 [15:08<00:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4553     0.1207        438        640:  96%|█████████▌| 775/809 [15:10<00:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105     0.4553     0.1206        195        640:  96%|█████████▌| 777/809 [15:13<00:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4555     0.1206        127        640:  96%|█████████▌| 778/809 [15:14<00:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4554     0.1205        219        640:  96%|█████████▋| 780/809 [15:16<00:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4553     0.1205        252        640:  97%|█████████▋| 782/809 [15:19<00:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.104     0.4552     0.1205        211        640:  97%|█████████▋| 783/809 [15:20<00:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.105      0.455     0.1205        367        640:  97%|█████████▋| 785/809 [15:22<00:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4546     0.1207        440        640:  97%|█████████▋| 787/809 [15:25<00:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4545     0.1207        151        640:  97%|█████████▋| 788/809 [15:26<00:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4545     0.1206        253        640:  98%|█████████▊| 790/809 [15:28<00:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4547     0.1205        265        640:  98%|█████████▊| 792/809 [15:31<00:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4546     0.1205        277        640:  98%|█████████▊| 794/809 [15:33<00:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4546     0.1205        431        640:  98%|█████████▊| 795/809 [15:34<00:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4545     0.1204        153        640:  99%|█████████▊| 797/809 [15:36<00:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4544     0.1203        109        640:  99%|█████████▉| 799/809 [15:39<00:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4544     0.1203        270        640:  99%|█████████▉| 800/809 [15:40<00:10,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.106     0.4544     0.1202        154        640:  99%|█████████▉| 802/809 [15:42<00:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4542     0.1203        407        640:  99%|█████████▉| 804/809 [15:45<00:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4542     0.1202        203        640: 100%|█████████▉| 805/809 [15:46<00:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     75/300      13.7G      1.107     0.4542     0.1202        262        640: 100%|█████████▉| 807/809 [15:48<00:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   1%|▏         | 1/69 [00:00<00:17,  3.86it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  14%|█▍        | 10/69 [00:02<00:14,  4.14it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  26%|██▌       | 18/69 [00:04<00:12,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  38%|███▊      | 26/69 [00:06<00:10,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  48%|████▊     | 33/69 [00:08<00:09,  3.71it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  61%|██████    | 42/69 [00:10<00:06,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  71%|███████   | 49/69 [00:12<00:05,  3.74it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  83%|████████▎ | 57/69 [00:14<00:02,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  94%|█████████▍| 65/69 [00:16<00:00,  4.07it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.064     0.4909    0.09943        213        640:   0%|          | 1/809 [00:01<15:34,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.263     0.4242     0.1134        327        640:   0%|          | 3/809 [00:03<15:26,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.38     0.3816     0.1305        399        640:   1%|          | 5/809 [00:05<15:33,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.344     0.3938     0.1228        198        640:   1%|          | 7/809 [00:08<15:24,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G        1.3     0.4176     0.1177        110        640:   1%|          | 8/809 [00:09<15:19,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.317     0.4095     0.1161        374        640:   1%|          | 10/809 [00:11<15:35,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.261     0.4204     0.1136        338        640:   1%|▏         | 12/809 [00:13<15:23,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.206     0.4361     0.1111        127        640:   2%|▏         | 13/809 [00:15<15:35,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.184     0.4423     0.1097         68        640:   2%|▏         | 15/809 [00:17<15:22,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.156     0.4481     0.1065        242        640:   2%|▏         | 17/809 [00:19<15:26,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4481     0.1033        165        640:   2%|▏         | 19/809 [00:22<15:28,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.168     0.4433     0.1052        365        640:   2%|▏         | 20/809 [00:23<15:24,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.183     0.4378     0.1087        240        640:   3%|▎         | 22/809 [00:25<15:32,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.185     0.4376     0.1084        248        640:   3%|▎         | 24/809 [00:27<15:13,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.185      0.437     0.1078        360        640:   3%|▎         | 25/809 [00:29<15:23,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.176     0.4407     0.1052        305        640:   3%|▎         | 27/809 [00:31<15:27,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.169     0.4433     0.1049        274        640:   4%|▎         | 29/809 [00:33<15:06,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.173     0.4428     0.1123        148        640:   4%|▍         | 31/809 [00:36<15:08,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.182     0.4398     0.1192        584        640:   4%|▍         | 32/809 [00:37<15:26,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.17     0.4425     0.1176        222        640:   4%|▍         | 34/809 [00:39<15:21,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.165     0.4444     0.1185        304        640:   4%|▍         | 36/809 [00:42<15:06,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.157     0.4477     0.1175        198        640:   5%|▍         | 37/809 [00:43<15:02,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.147     0.4503     0.1162        157        640:   5%|▍         | 39/809 [00:45<14:48,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.141     0.4526     0.1154        249        640:   5%|▌         | 41/809 [00:47<14:43,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.14     0.4524     0.1149        254        640:   5%|▌         | 42/809 [00:49<15:15,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.15     0.4483      0.115        344        640:   5%|▌         | 44/809 [00:51<15:19,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.138       0.45      0.113        141        640:   6%|▌         | 46/809 [00:53<14:56,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.134     0.4507     0.1122        255        640:   6%|▌         | 47/809 [00:55<14:46,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.141     0.4489      0.116        664        640:   6%|▌         | 49/809 [00:57<15:27,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145      0.446     0.1148        234        640:   6%|▋         | 51/809 [01:00<15:14,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.141     0.4463     0.1145        207        640:   6%|▋         | 52/809 [01:01<15:34,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.157     0.4417     0.1223        618        640:   7%|▋         | 54/809 [01:03<15:33,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.152     0.4435     0.1215        201        640:   7%|▋         | 56/809 [01:06<14:47,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.149     0.4445     0.1209        199        640:   7%|▋         | 57/809 [01:07<14:40,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.137     0.4481     0.1193        158        640:   7%|▋         | 59/809 [01:09<14:44,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145     0.4458     0.1203        247        640:   8%|▊         | 61/809 [01:11<14:34,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.148     0.4448     0.1213        377        640:   8%|▊         | 63/809 [01:14<14:25,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.152      0.443     0.1211        351        640:   8%|▊         | 64/809 [01:15<14:27,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4412     0.1209        181        640:   8%|▊         | 66/809 [01:17<14:36,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4425       0.12        392        640:   8%|▊         | 68/809 [01:20<14:21,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.143      0.444     0.1189        140        640:   9%|▊         | 70/809 [01:22<14:13,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.142     0.4442     0.1186        309        640:   9%|▉         | 71/809 [01:23<14:37,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.138     0.4455     0.1179        303        640:   9%|▉         | 73/809 [01:25<14:12,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.136     0.4459     0.1177        150        640:   9%|▉         | 75/809 [01:28<14:14,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.133     0.4464     0.1175        179        640:   9%|▉         | 76/809 [01:29<14:17,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.128      0.447     0.1168        283        640:  10%|▉         | 78/809 [01:31<14:08,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.131     0.4449     0.1167        364        640:  10%|▉         | 80/809 [01:34<14:17,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4464     0.1162        131        640:  10%|█         | 81/809 [01:35<14:04,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118      0.448     0.1158        204        640:  10%|█         | 83/809 [01:37<14:09,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4483     0.1164        232        640:  11%|█         | 85/809 [01:39<13:54,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4493      0.118        145        640:  11%|█         | 87/809 [01:42<14:17,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4492     0.1178        198        640:  11%|█         | 88/809 [01:43<14:31,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4481     0.1184        414        640:  11%|█         | 90/809 [01:45<14:16,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.122     0.4481     0.1182        179        640:  11%|█▏        | 92/809 [01:48<14:05,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4495     0.1179         92        640:  11%|█▏        | 93/809 [01:49<14:00,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4481     0.1185        333        640:  12%|█▏        | 95/809 [01:51<13:48,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.123     0.4489      0.118        183        640:  12%|█▏        | 97/809 [01:54<13:48,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4482     0.1178        278        640:  12%|█▏        | 99/809 [01:56<13:59,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.127     0.4474     0.1182        289        640:  12%|█▏        | 100/809 [01:57<13:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4489     0.1176        159        640:  13%|█▎        | 102/809 [01:59<13:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.125     0.4489     0.1177        354        640:  13%|█▎        | 104/809 [02:02<13:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.123     0.4497     0.1173        239        640:  13%|█▎        | 105/809 [02:03<13:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.122     0.4502     0.1171        253        640:  13%|█▎        | 107/809 [02:05<13:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.13     0.4481     0.1193        574        640:  13%|█▎        | 109/809 [02:08<13:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.127     0.4488     0.1187        346        640:  14%|█▎        | 111/809 [02:10<13:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.127     0.4486     0.1187        301        640:  14%|█▍        | 112/809 [02:11<13:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.123     0.4498     0.1179        214        640:  14%|█▍        | 114/809 [02:14<13:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.123     0.4496     0.1178        181        640:  14%|█▍        | 116/809 [02:16<13:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.126     0.4484     0.1184        375        640:  14%|█▍        | 117/809 [02:17<13:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124      0.449     0.1186        173        640:  15%|█▍        | 119/809 [02:19<13:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4487     0.1184        344        640:  15%|█▍        | 121/809 [02:22<13:07,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4498     0.1181         69        640:  15%|█▌        | 123/809 [02:24<13:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12      0.451     0.1178        118        640:  15%|█▌        | 124/809 [02:25<13:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4511      0.118        122        640:  16%|█▌        | 126/809 [02:27<13:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4508     0.1186        171        640:  16%|█▌        | 128/809 [02:30<13:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.126     0.4508     0.1187        148        640:  16%|█▌        | 130/809 [02:32<13:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4517     0.1186        112        640:  16%|█▋        | 132/809 [02:35<13:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.126     0.4513     0.1191        266        640:  16%|█▋        | 133/809 [02:36<13:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.125     0.4516     0.1187        291        640:  17%|█▋        | 135/809 [02:38<13:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.125     0.4521     0.1186        477        640:  17%|█▋        | 137/809 [02:40<12:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.127     0.4513     0.1187        328        640:  17%|█▋        | 138/809 [02:42<13:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.131     0.4505     0.1189        233        640:  17%|█▋        | 140/809 [02:44<13:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.138     0.4487     0.1209        125        640:  18%|█▊        | 142/809 [02:46<13:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.141     0.4477     0.1213        403        640:  18%|█▊        | 143/809 [02:48<13:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.147     0.4462     0.1218        368        640:  18%|█▊        | 145/809 [02:50<12:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.143     0.4471     0.1214        225        640:  18%|█▊        | 147/809 [02:52<13:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.143     0.4471     0.1211        369        640:  18%|█▊        | 149/809 [02:55<12:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.141     0.4477     0.1209        195        640:  19%|█▊        | 150/809 [02:56<12:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.138     0.4482     0.1205        257        640:  19%|█▉        | 152/809 [02:58<12:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.142     0.4472     0.1221        351        640:  19%|█▉        | 154/809 [03:01<13:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.141     0.4474     0.1218        213        640:  19%|█▉        | 155/809 [03:02<13:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.142     0.4467     0.1221        370        640:  19%|█▉        | 157/809 [03:04<12:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.146     0.4456     0.1224        463        640:  20%|█▉        | 159/809 [03:07<12:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145     0.4455     0.1222        205        640:  20%|█▉        | 160/809 [03:08<12:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.144     0.4458     0.1218        116        640:  20%|██        | 162/809 [03:10<12:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145     0.4459     0.1222        308        640:  20%|██        | 164/809 [03:12<12:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145      0.446      0.122        201        640:  20%|██        | 165/809 [03:14<13:02,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.142      0.447     0.1217        202        640:  21%|██        | 167/809 [03:16<12:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.141     0.4473     0.1214        152        640:  21%|██        | 169/809 [03:18<12:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.144      0.446     0.1215        326        640:  21%|██        | 171/809 [03:21<12:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.143     0.4466     0.1214        103        640:  21%|██▏       | 172/809 [03:22<12:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.144     0.4466     0.1215        249        640:  22%|██▏       | 174/809 [03:24<12:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.146     0.4468     0.1227        170        640:  22%|██▏       | 176/809 [03:27<12:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.149     0.4462     0.1235        304        640:  22%|██▏       | 177/809 [03:28<12:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.149     0.4459     0.1235        353        640:  22%|██▏       | 179/809 [03:30<12:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.15     0.4455     0.1234        238        640:  22%|██▏       | 181/809 [03:33<12:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.15     0.4458     0.1232         99        640:  22%|██▏       | 182/809 [03:34<12:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.149     0.4458     0.1231        114        640:  23%|██▎       | 184/809 [03:36<12:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4442     0.1238        355        640:  23%|██▎       | 186/809 [03:39<12:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.156     0.4436      0.124        176        640:  23%|██▎       | 188/809 [03:41<12:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.159     0.4426      0.125        330        640:  23%|██▎       | 189/809 [03:42<12:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.159     0.4423     0.1247        160        640:  24%|██▎       | 191/809 [03:44<12:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.16     0.4424     0.1248        117        640:  24%|██▍       | 193/809 [03:47<12:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.161     0.4423     0.1248        265        640:  24%|██▍       | 194/809 [03:48<12:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.163     0.4416     0.1247        341        640:  24%|██▍       | 196/809 [03:50<12:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.162     0.4416     0.1244        285        640:  24%|██▍       | 198/809 [03:53<11:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.163     0.4413     0.1243        292        640:  25%|██▍       | 199/809 [03:54<11:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.162     0.4413     0.1243        121        640:  25%|██▍       | 201/809 [03:56<11:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.161     0.4415      0.124        254        640:  25%|██▌       | 203/809 [03:59<11:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.162     0.4409     0.1237        239        640:  25%|██▌       | 205/809 [04:01<11:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.162     0.4408     0.1236        287        640:  25%|██▌       | 206/809 [04:02<11:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.163     0.4403     0.1237        417        640:  26%|██▌       | 208/809 [04:04<11:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.163     0.4399     0.1233        205        640:  26%|██▌       | 210/809 [04:07<11:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.163     0.4397     0.1235        296        640:  26%|██▌       | 211/809 [04:08<11:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.164     0.4394     0.1234        258        640:  26%|██▋       | 213/809 [04:10<11:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.165     0.4388     0.1234        336        640:  27%|██▋       | 215/809 [04:12<11:16,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.166     0.4387     0.1231        329        640:  27%|██▋       | 217/809 [04:15<11:18,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.166     0.4384     0.1232        381        640:  27%|██▋       | 218/809 [04:16<11:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.165     0.4386     0.1231        146        640:  27%|██▋       | 220/809 [04:18<11:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.165     0.4386     0.1231        385        640:  27%|██▋       | 222/809 [04:21<11:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.164     0.4384     0.1231        165        640:  28%|██▊       | 223/809 [04:22<11:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.163     0.4385     0.1228        158        640:  28%|██▊       | 225/809 [04:24<11:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.161      0.439     0.1225        202        640:  28%|██▊       | 227/809 [04:27<11:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.163     0.4385     0.1235        421        640:  28%|██▊       | 228/809 [04:28<11:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.162     0.4384     0.1233        406        640:  28%|██▊       | 230/809 [04:30<11:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.161     0.4385     0.1232        214        640:  29%|██▊       | 232/809 [04:33<11:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.16     0.4387     0.1231        182        640:  29%|██▉       | 233/809 [04:34<11:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.158     0.4395     0.1227        215        640:  29%|██▉       | 235/809 [04:36<11:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.16     0.4385      0.123        211        640:  29%|██▉       | 237/809 [04:38<11:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.16     0.4388     0.1229        222        640:  30%|██▉       | 239/809 [04:41<10:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.159      0.439     0.1228        392        640:  30%|██▉       | 240/809 [04:42<10:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.157     0.4395     0.1224        185        640:  30%|██▉       | 242/809 [04:44<11:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.155       0.44      0.122        147        640:  30%|███       | 244/809 [04:47<10:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.153     0.4403     0.1219        126        640:  30%|███       | 246/809 [04:49<10:48,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.153     0.4403     0.1219        175        640:  31%|███       | 247/809 [04:50<10:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4407     0.1216        232        640:  31%|███       | 249/809 [04:52<10:43,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.149     0.4412     0.1214         92        640:  31%|███       | 251/809 [04:55<10:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4404     0.1217        125        640:  31%|███▏      | 253/809 [04:57<10:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.152     0.4403     0.1217        272        640:  31%|███▏      | 254/809 [04:58<10:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.153     0.4397     0.1217        273        640:  32%|███▏      | 256/809 [05:00<10:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.153     0.4397     0.1216        180        640:  32%|███▏      | 258/809 [05:03<10:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.153     0.4397     0.1215        316        640:  32%|███▏      | 259/809 [05:04<10:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.153     0.4395     0.1216        497        640:  32%|███▏      | 261/809 [05:06<10:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4391     0.1221        121        640:  33%|███▎      | 263/809 [05:09<10:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.155     0.4389      0.122        189        640:  33%|███▎      | 264/809 [05:10<10:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.153     0.4393     0.1218        126        640:  33%|███▎      | 266/809 [05:12<10:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.153     0.4396     0.1217        231        640:  33%|███▎      | 268/809 [05:15<10:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4393     0.1216        239        640:  33%|███▎      | 270/809 [05:17<10:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.152     0.4398     0.1216        130        640:  33%|███▎      | 271/809 [05:18<10:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4396     0.1222        177        640:  34%|███▎      | 273/809 [05:20<10:18,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4393     0.1221        195        640:  34%|███▍      | 275/809 [05:23<10:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4392     0.1221        281        640:  34%|███▍      | 276/809 [05:24<10:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.153     0.4394      0.122         84        640:  34%|███▍      | 278/809 [05:26<10:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.152     0.4396     0.1217        165        640:  35%|███▍      | 280/809 [05:29<10:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.152     0.4393     0.1218        165        640:  35%|███▍      | 282/809 [05:31<10:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4394     0.1216        239        640:  35%|███▌      | 284/809 [05:33<10:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4395     0.1216        270        640:  35%|███▌      | 285/809 [05:35<10:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.15     0.4398     0.1215        351        640:  35%|███▌      | 287/809 [05:37<10:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4395     0.1223        174        640:  36%|███▌      | 289/809 [05:39<10:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.15     0.4396     0.1222        238        640:  36%|███▌      | 290/809 [05:41<10:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.15     0.4396     0.1221        185        640:  36%|███▌      | 292/809 [05:43<10:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.149       0.44      0.122        306        640:  36%|███▋      | 294/809 [05:45<10:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.146     0.4408     0.1219        154        640:  37%|███▋      | 296/809 [05:48<10:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.146     0.4409     0.1219        187        640:  37%|███▋      | 297/809 [05:49<09:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.143     0.4415     0.1217        194        640:  37%|███▋      | 299/809 [05:51<10:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145     0.4412      0.122        255        640:  37%|███▋      | 301/809 [05:54<10:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.146     0.4409     0.1222        382        640:  37%|███▋      | 302/809 [05:55<10:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.144     0.4411      0.122        194        640:  38%|███▊      | 304/809 [05:57<10:22,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.143     0.4415     0.1219        100        640:  38%|███▊      | 305/809 [05:59<10:23,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.142     0.4419     0.1218        191        640:  38%|███▊      | 307/809 [06:01<10:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.143     0.4419     0.1219        166        640:  38%|███▊      | 309/809 [06:03<10:21,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.143     0.4418     0.1219        398        640:  38%|███▊      | 310/809 [06:05<10:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.142     0.4422     0.1216        145        640:  39%|███▊      | 312/809 [06:07<09:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145     0.4413     0.1226        485        640:  39%|███▉      | 314/809 [06:10<10:26,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145     0.4413     0.1226        307        640:  39%|███▉      | 315/809 [06:11<10:06,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.148     0.4406     0.1236        347        640:  39%|███▉      | 317/809 [06:13<09:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.149     0.4401     0.1238        540        640:  39%|███▉      | 319/809 [06:15<09:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.149     0.4401     0.1238        281        640:  40%|███▉      | 321/809 [06:18<09:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.149     0.4402     0.1237        226        640:  40%|███▉      | 322/809 [06:19<09:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.15       0.44     0.1238        424        640:  40%|████      | 324/809 [06:21<09:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4394     0.1242        395        640:  40%|████      | 326/809 [06:23<09:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4396     0.1241        205        640:  40%|████      | 327/809 [06:25<09:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4387     0.1251        326        640:  41%|████      | 329/809 [06:27<09:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.153     0.4391     0.1249        186        640:  41%|████      | 331/809 [06:29<09:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4387     0.1259        475        640:  41%|████      | 333/809 [06:32<09:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.155     0.4385     0.1258        310        640:  41%|████▏     | 334/809 [06:33<09:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.155     0.4386     0.1257        146        640:  42%|████▏     | 336/809 [06:35<09:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.156     0.4381     0.1263        392        640:  42%|████▏     | 338/809 [06:38<09:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.156     0.4383     0.1263        219        640:  42%|████▏     | 339/809 [06:39<09:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.155     0.4385     0.1262        313        640:  42%|████▏     | 341/809 [06:41<09:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.155     0.4387     0.1261        218        640:  42%|████▏     | 343/809 [06:44<09:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4389      0.126         87        640:  43%|████▎     | 344/809 [06:45<09:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.155     0.4386     0.1261        329        640:  43%|████▎     | 346/809 [06:47<09:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.155     0.4386      0.126        114        640:  43%|████▎     | 348/809 [06:49<09:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4389     0.1259        130        640:  43%|████▎     | 350/809 [06:52<08:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154      0.439     0.1258        158        640:  43%|████▎     | 351/809 [06:53<08:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4387     0.1259        276        640:  44%|████▎     | 353/809 [06:55<08:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.155     0.4386     0.1258        389        640:  44%|████▍     | 355/809 [06:58<08:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.154     0.4386     0.1256        185        640:  44%|████▍     | 356/809 [06:59<08:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.155     0.4383     0.1257        259        640:  44%|████▍     | 358/809 [07:01<08:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.153     0.4389     0.1254        117        640:  44%|████▍     | 360/809 [07:04<08:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.152     0.4391     0.1253         99        640:  45%|████▍     | 361/809 [07:05<09:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.152     0.4393     0.1252        117        640:  45%|████▍     | 363/809 [07:07<08:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4395      0.125        178        640:  45%|████▌     | 365/809 [07:09<08:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4395     0.1249        205        640:  45%|████▌     | 367/809 [07:12<08:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.151     0.4394     0.1248        286        640:  45%|████▌     | 368/809 [07:13<08:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.15     0.4397     0.1247        185        640:  46%|████▌     | 370/809 [07:16<08:54,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.149     0.4398     0.1245        214        640:  46%|████▌     | 372/809 [07:18<08:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.148       0.44     0.1245        177        640:  46%|████▌     | 373/809 [07:19<08:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.148     0.4403     0.1243        263        640:  46%|████▋     | 375/809 [07:21<08:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.146     0.4406     0.1242        419        640:  47%|████▋     | 377/809 [07:24<08:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145     0.4408     0.1241        215        640:  47%|████▋     | 378/809 [07:25<08:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.146     0.4407      0.124        282        640:  47%|████▋     | 380/809 [07:27<08:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.146     0.4405      0.124        187        640:  47%|████▋     | 382/809 [07:30<08:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.146     0.4406      0.124        236        640:  47%|████▋     | 384/809 [07:32<08:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145      0.441     0.1239        159        640:  48%|████▊     | 385/809 [07:33<08:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.147     0.4405     0.1244        223        640:  48%|████▊     | 387/809 [07:35<08:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.146     0.4408     0.1241        220        640:  48%|████▊     | 389/809 [07:38<08:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.146     0.4407     0.1243        375        640:  48%|████▊     | 390/809 [07:39<08:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145     0.4411     0.1241        392        640:  48%|████▊     | 392/809 [07:41<08:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145     0.4412     0.1239        216        640:  49%|████▊     | 394/809 [07:44<08:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.145     0.4411      0.124        397        640:  49%|████▉     | 395/809 [07:45<08:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.144     0.4416     0.1238        150        640:  49%|████▉     | 397/809 [07:47<08:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.142     0.4421     0.1236        366        640:  49%|████▉     | 399/809 [07:50<07:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.142     0.4424     0.1235        342        640:  50%|████▉     | 401/809 [07:52<07:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.142     0.4423     0.1234        157        640:  50%|████▉     | 402/809 [07:53<07:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.141     0.4427     0.1233        281        640:  50%|████▉     | 404/809 [07:55<07:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.141     0.4428     0.1232        299        640:  50%|█████     | 406/809 [07:58<07:45,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.14     0.4433      0.123        139        640:  50%|█████     | 408/809 [08:00<07:42,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.139     0.4435     0.1229        181        640:  51%|█████     | 409/809 [08:01<07:39,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.138     0.4438     0.1228        203        640:  51%|█████     | 411/809 [08:03<07:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.138     0.4437     0.1227        318        640:  51%|█████     | 413/809 [08:06<07:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.138     0.4434     0.1227        335        640:  51%|█████     | 414/809 [08:07<07:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.138     0.4437     0.1226        119        640:  51%|█████▏    | 416/809 [08:09<07:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.137     0.4436     0.1225        151        640:  52%|█████▏    | 418/809 [08:12<07:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.137     0.4436     0.1224        225        640:  52%|█████▏    | 420/809 [08:14<07:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.137     0.4436     0.1225        359        640:  52%|█████▏    | 421/809 [08:15<07:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.137     0.4436     0.1223        219        640:  52%|█████▏    | 423/809 [08:17<07:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.135     0.4439     0.1222         78        640:  53%|█████▎    | 425/809 [08:20<07:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.135     0.4439     0.1221        213        640:  53%|█████▎    | 426/809 [08:21<07:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.135     0.4439      0.122        171        640:  53%|█████▎    | 428/809 [08:23<07:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.135     0.4438      0.122        155        640:  53%|█████▎    | 430/809 [08:26<07:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.134     0.4439     0.1218        155        640:  53%|█████▎    | 432/809 [08:28<07:14,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.134     0.4438     0.1218        210        640:  54%|█████▎    | 434/809 [08:31<07:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.134     0.4438     0.1218        458        640:  54%|█████▍    | 435/809 [08:32<07:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.134     0.4441     0.1217        375        640:  54%|█████▍    | 437/809 [08:34<07:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.134     0.4441     0.1217        539        640:  54%|█████▍    | 439/809 [08:36<07:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.134     0.4441     0.1217        263        640:  54%|█████▍    | 440/809 [08:38<07:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.133     0.4443     0.1215        197        640:  55%|█████▍    | 442/809 [08:40<07:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.134     0.4442     0.1215        182        640:  55%|█████▍    | 444/809 [08:42<07:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.134     0.4441     0.1215        201        640:  55%|█████▌    | 446/809 [08:44<06:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.135      0.444     0.1216        456        640:  55%|█████▌    | 447/809 [08:46<06:57,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.135     0.4437     0.1216        258        640:  56%|█████▌    | 449/809 [08:48<06:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.135     0.4439     0.1215        275        640:  56%|█████▌    | 451/809 [08:50<07:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.134      0.444     0.1214        298        640:  56%|█████▌    | 452/809 [08:52<07:15,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.133     0.4445     0.1212        171        640:  56%|█████▌    | 454/809 [08:54<07:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.133     0.4446     0.1211        221        640:  56%|█████▋    | 456/809 [08:56<06:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.132      0.445     0.1211        132        640:  56%|█████▋    | 457/809 [08:58<06:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.132     0.4451      0.121        375        640:  57%|█████▋    | 459/809 [09:00<06:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.131     0.4454     0.1208        297        640:  57%|█��███▋    | 461/809 [09:02<06:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.131     0.4454     0.1208        456        640:  57%|█████▋    | 463/809 [09:05<06:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.131     0.4455     0.1207        203        640:  57%|█████▋    | 464/809 [09:06<06:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.131     0.4454     0.1209        436        640:  58%|█████▊    | 466/809 [09:08<06:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.131     0.4453     0.1209        356        640:  58%|█████▊    | 468/809 [09:10<06:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.131     0.4455     0.1208        189        640:  58%|█████▊    | 469/809 [09:12<06:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.129     0.4456     0.1207        231        640:  58%|█████▊    | 471/809 [09:14<06:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.129     0.4458     0.1208        317        640:  58%|█████▊    | 473/809 [09:16<06:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.128     0.4463     0.1206        198        640:  59%|█████▊    | 475/809 [09:19<06:25,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.128     0.4462     0.1207        193        640:  59%|█████▉    | 476/809 [09:20<06:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.128     0.4463     0.1207        250        640:  59%|█████▉    | 478/809 [09:22<06:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.128     0.4461     0.1208        747        640:  59%|█████▉    | 480/809 [09:25<06:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.128     0.4464     0.1207        190        640:  59%|█████▉    | 481/809 [09:26<06:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.127     0.4465     0.1206        247        640:  60%|█████▉    | 483/809 [09:28<06:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.126      0.447     0.1205        206        640:  60%|█████▉    | 485/809 [09:31<06:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.127     0.4468     0.1207        461        640:  60%|██████    | 486/809 [09:32<06:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.128     0.4465     0.1207        551        640:  60%|██████    | 488/809 [09:34<06:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.127     0.4468     0.1207        267        640:  61%|██████    | 490/809 [09:37<06:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.127     0.4469     0.1207        348        640:  61%|██████    | 491/809 [09:38<06:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.127     0.4469      0.121        162        640:  61%|██████    | 493/809 [09:40<06:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.126     0.4472     0.1209        253        640:  61%|██████    | 495/809 [09:42<06:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.125     0.4476     0.1207        102        640:  61%|██████▏   | 497/809 [09:45<06:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.125     0.4477     0.1207        241        640:  62%|██████▏   | 498/809 [09:46<06:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4479     0.1206        238        640:  62%|██████▏   | 500/809 [09:48<06:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.123     0.4481     0.1205        150        640:  62%|██████▏   | 502/809 [09:51<05:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.122     0.4483     0.1204        158        640:  62%|██████▏   | 503/809 [09:52<05:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4487     0.1202        131        640:  62%|██████▏   | 505/809 [09:54<05:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.122     0.4486     0.1203        248        640:  63%|██████▎   | 507/809 [09:56<05:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.122     0.4485     0.1203        248        640:  63%|██████▎   | 509/809 [09:59<05:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.122     0.4486     0.1202        154        640:  63%|██████▎   | 510/809 [10:00<05:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4488     0.1201        123        640:  63%|██████▎   | 512/809 [10:02<05:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4489     0.1199        157        640:  64%|██████▎   | 514/809 [10:05<05:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12      0.449     0.1199        257        640:  64%|██████▎   | 515/809 [10:06<05:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4485     0.1201        387        640:  64%|██████▍   | 517/809 [10:08<05:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4488       0.12        249        640:  64%|██████▍   | 519/809 [10:11<05:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4492     0.1199        105        640:  64%|██████▍   | 520/809 [10:13<05:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4494     0.1198         77        640:  65%|██████▍   | 522/809 [10:14<05:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4493     0.1197        276        640:  65%|██████▍   | 524/809 [10:16<05:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4498     0.1196        130        640:  65%|██████▌   | 526/809 [10:19<05:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4499     0.1195        233        640:  65%|██████▌   | 527/809 [10:20<05:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4496     0.1199        270        640:  65%|██████▌   | 529/809 [10:22<05:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4495     0.1198        253        640:  66%|██████▌   | 531/809 [10:25<05:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4495     0.1198        213        640:  66%|██████▌   | 532/809 [10:26<05:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4496     0.1196        294        640:  66%|██████▌   | 534/809 [10:28<05:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4497     0.1195        315        640:  66%|██████▋   | 536/809 [10:31<05:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119       0.45     0.1194        252        640:  67%|██████▋   | 538/809 [10:33<05:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4498     0.1195        283        640:  67%|██████▋   | 539/809 [10:34<05:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4499     0.1194        349        640:  67%|██████▋   | 541/809 [10:36<05:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4499     0.1193        116        640:  67%|██████▋   | 543/809 [10:39<05:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4499     0.1196        345        640:  67%|██████▋   | 545/809 [10:41<05:04,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4497     0.1196        467        640:  67%|██████▋   | 546/809 [10:42<05:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4498     0.1195        278        640:  68%|██████▊   | 548/809 [10:44<05:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4497     0.1195        444        640:  68%|██████▊   | 550/809 [10:47<04:59,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4496     0.1195        202        640:  68%|██████▊   | 551/809 [10:48<04:56,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4494     0.1194        411        640:  68%|██████▊   | 553/809 [10:50<04:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4497     0.1193        154        640:  69%|██████▊   | 555/809 [10:53<04:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4499     0.1192        184        640:  69%|██████▉   | 557/809 [10:55<04:49,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4501     0.1192        167        640:  69%|██████▉   | 558/809 [10:56<04:49,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4502     0.1191        266        640:  69%|██████▉   | 560/809 [10:58<04:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4504     0.1189        226        640:  69%|██████▉   | 562/809 [11:01<04:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4504     0.1188        251        640:  70%|██████▉   | 564/809 [11:03<04:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4504     0.1188        283        640:  70%|██████▉   | 565/809 [11:04<04:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4502     0.1188        227        640:  70%|███████   | 567/809 [11:07<04:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4499     0.1188        411        640:  70%|███████   | 569/809 [11:09<04:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4498     0.1188        227        640:  70%|███████   | 570/809 [11:10<04:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4498     0.1188        280        640:  71%|███████   | 572/809 [11:13<04:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4496     0.1189        338        640:  71%|███████   | 574/809 [11:15<04:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4496     0.1189        227        640:  71%|███████   | 575/809 [11:16<04:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4497     0.1188        131        640:  71%|███████▏  | 577/809 [11:19<04:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4496     0.1188        490        640:  72%|███████▏  | 579/809 [11:21<04:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4496     0.1188        280        640:  72%|███████▏  | 580/809 [11:22<04:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4495     0.1189        370        640:  72%|███████▏  | 582/809 [11:24<04:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4492      0.119        422        640:  72%|███████▏  | 584/809 [11:27<04:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12      0.449     0.1191        321        640:  72%|███████▏  | 586/809 [11:29<04:33,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4489     0.1192        461        640:  73%|███████▎  | 587/809 [11:31<04:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4489     0.1191        390        640:  73%|███████▎  | 589/809 [11:33<04:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4491      0.119        125        640:  73%|███████▎  | 591/809 [11:35<04:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.122     0.4488     0.1195        625        640:  73%|███████▎  | 592/809 [11:37<04:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4492     0.1194         84        640:  73%|███████▎  | 594/809 [11:39<04:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4493     0.1194        226        640:  74%|███████▎  | 596/809 [11:41<04:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4496     0.1193         98        640:  74%|███████▍  | 598/809 [11:43<04:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4497     0.1192        319        640:  74%|███████▍  | 599/809 [11:45<04:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4497     0.1192        209        640:  74%|███████▍  | 601/809 [11:47<04:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4498     0.1191        294        640:  75%|███████▍  | 603/809 [11:49<03:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4497     0.1192        725        640:  75%|███████▍  | 604/809 [11:51<04:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4499     0.1191        345        640:  75%|███████▍  | 606/809 [11:53<03:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119       0.45     0.1191        388        640:  75%|███████▌  | 608/809 [11:55<03:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4504      0.119         93        640:  75%|███████▌  | 610/809 [11:58<03:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4506      0.119        452        640:  76%|███████▌  | 611/809 [11:59<03:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117      0.451     0.1189        224        640:  76%|███████▌  | 613/809 [12:01<03:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4511      0.119        433        640:  76%|███████▌  | 615/809 [12:04<03:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4512      0.119        246        640:  76%|███████▌  | 616/809 [12:05<03:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4513     0.1193        128        640:  76%|███████▋  | 618/809 [12:07<03:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4513     0.1193        437        640:  77%|███████▋  | 620/809 [12:09<03:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4516     0.1193        169        640:  77%|███████▋  | 621/809 [12:11<03:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4516     0.1192        274        640:  77%|███████▋  | 623/809 [12:13<03:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4516     0.1192        240        640:  77%|███████▋  | 625/809 [12:15<03:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4516     0.1193        385        640:  78%|███████▊  | 627/809 [12:18<03:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4516     0.1193        286        640:  78%|███████▊  | 628/809 [12:19<03:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4517     0.1192        178        640:  78%|███████▊  | 630/809 [12:21<03:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116      0.452      0.119        245        640:  78%|███████▊  | 632/809 [12:23<03:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4519     0.1191        454        640:  78%|███████▊  | 633/809 [12:25<03:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.115      0.452      0.119         84        640:  78%|███████▊  | 635/809 [12:27<03:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116      0.452     0.1189        251        640:  79%|███████▊  | 637/809 [12:29<03:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4519     0.1188        173        640:  79%|███████▉  | 639/809 [12:32<03:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4519     0.1188        146        640:  79%|███████▉  | 640/809 [12:33<03:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4517     0.1188        218        640:  79%|███████▉  | 642/809 [12:35<03:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4519     0.1187        255        640:  80%|███████▉  | 644/809 [12:38<03:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116      0.452     0.1186        127        640:  80%|███████▉  | 645/809 [12:39<03:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.115     0.4522     0.1185        177        640:  80%|███████▉  | 647/809 [12:41<03:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4522     0.1186        170        640:  80%|████████  | 649/809 [12:44<03:16,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4522     0.1185        280        640:  80%|████████  | 650/809 [12:45<03:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4521     0.1185        148        640:  81%|████████  | 652/809 [12:47<03:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.115     0.4524     0.1185        187        640:  81%|████████  | 654/809 [12:50<03:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.115     0.4522     0.1185        209        640:  81%|████████  | 655/809 [12:51<02:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4523     0.1184        131        640:  81%|████████  | 657/809 [12:53<02:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4522     0.1183        221        640:  81%|████████▏ | 659/809 [12:55<02:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117      0.452     0.1188        397        640:  82%|████████▏ | 661/809 [12:58<02:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4519     0.1188        406        640:  82%|████████▏ | 662/809 [12:59<02:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521     0.1188        252        640:  82%|████████▏ | 664/809 [13:01<02:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4522     0.1188        205        640:  82%|████████▏ | 666/809 [13:04<02:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117      0.452     0.1189        360        640:  82%|████████▏ | 667/809 [13:05<02:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4519     0.1191        194        640:  83%|████████▎ | 669/809 [13:07<02:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4518     0.1191        192        640:  83%|████████▎ | 671/809 [13:10<02:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4518     0.1191        417        640:  83%|████████▎ | 672/809 [13:11<02:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4519      0.119        367        640:  83%|████████▎ | 674/809 [13:13<02:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521      0.119        156        640:  84%|████████▎ | 676/809 [13:16<02:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521     0.1189        160        640:  84%|████████▎ | 677/809 [13:17<02:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117      0.452     0.1191         89        640:  84%|████████▍ | 679/809 [13:19<02:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118      0.452      0.119        157        640:  84%|████████▍ | 681/809 [13:21<02:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521      0.119        254        640:  84%|████████▍ | 683/809 [13:24<02:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117      0.452     0.1195        271        640:  85%|████████▍ | 684/809 [13:25<02:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4522     0.1194        181        640:  85%|████████▍ | 686/809 [13:27<02:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.116     0.4523     0.1193        289        640:  85%|████████▌ | 688/809 [13:30<02:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521     0.1192        317        640:  85%|████████▌ | 690/809 [13:32<02:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521     0.1192        128        640:  85%|████████▌ | 691/809 [13:33<02:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521     0.1191        210        640:  86%|████████▌ | 693/809 [13:35<02:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521     0.1191        303        640:  86%|████████▌ | 695/809 [13:38<02:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4523      0.119        121        640:  86%|████████▌ | 697/809 [13:40<02:07,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4522      0.119        304        640:  86%|████████▋ | 698/809 [13:41<02:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4523      0.119        278        640:  87%|████████▋ | 700/809 [13:43<02:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521      0.119        381        640:  87%|████████▋ | 702/809 [13:46<02:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521      0.119        285        640:  87%|████████▋ | 703/809 [13:47<02:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117      0.452     0.1189        265        640:  87%|████████▋ | 705/809 [13:49<02:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521     0.1189        139        640:  87%|████████▋ | 707/809 [13:52<02:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521     0.1188        167        640:  88%|████████▊ | 708/809 [13:53<01:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4519     0.1189        223        640:  88%|████████▊ | 710/809 [13:55<01:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4518     0.1189        271        640:  88%|████████▊ | 712/809 [13:58<01:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4517     0.1189        340        640:  88%|████████▊ | 714/809 [14:00<01:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4516      0.119        420        640:  88%|████████▊ | 715/809 [14:01<01:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4514      0.119        263        640:  89%|████████▊ | 717/809 [14:03<01:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4513     0.1191        404        640:  89%|████████▉ | 719/809 [14:06<01:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4514     0.1192        338        640:  89%|████████▉ | 720/809 [14:07<01:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4513     0.1192        252        640:  89%|████████▉ | 722/809 [14:09<01:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4509       0.12        605        640:  89%|████████▉ | 724/809 [14:12<01:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.122     0.4507       0.12        362        640:  90%|████████▉ | 725/809 [14:13<01:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.122     0.4506       0.12        221        640:  90%|████████▉ | 727/809 [14:15<01:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.122     0.4506     0.1199        165        640:  90%|█████████ | 729/809 [14:18<01:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.123     0.4505       0.12        317        640:  90%|█████████ | 731/809 [14:20<01:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.123     0.4504     0.1199        284        640:  91%|█████████ | 733/809 [14:22<01:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.123     0.4505     0.1199        239        640:  91%|█████████ | 734/809 [14:24<01:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.123     0.4503     0.1199        498        640:  91%|█████████ | 736/809 [14:26<01:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4501     0.1202        217        640:  91%|█████████ | 738/809 [14:28<01:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4501     0.1202        182        640:  91%|█████████▏| 739/809 [14:30<01:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4502     0.1201        115        640:  92%|█████████▏| 741/809 [14:32<01:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4502     0.1201        211        640:  92%|█████████▏| 743/809 [14:34<01:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4501     0.1201        378        640:  92%|█████████▏| 745/809 [14:37<01:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4501     0.1201        277        640:  92%|█████████▏| 746/809 [14:38<01:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.124     0.4504     0.1202        209        640:  92%|█████████▏| 748/809 [14:40<01:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.123     0.4506       0.12        165        640:  93%|█████████▎| 750/809 [14:42<01:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.122     0.4507       0.12        139        640:  93%|█████████▎| 751/809 [14:44<01:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121      0.451     0.1199        182        640:  93%|█████████▎| 753/809 [14:46<01:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4511     0.1198        184        640:  93%|█████████▎| 755/809 [14:48<01:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4513     0.1197        102        640:  94%|█████████▎| 757/809 [14:51<01:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4513     0.1197         91        640:  94%|█████████▎| 758/809 [14:52<00:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4517     0.1195         81        640:  94%|█████████▍| 760/809 [14:54<00:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4518     0.1195         94        640:  94%|█████████▍| 762/809 [14:57<00:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4519     0.1194        242        640:  94%|█████████▍| 763/809 [14:58<00:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117      0.452     0.1194        348        640:  95%|█████████▍| 765/809 [15:00<00:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117      0.452     0.1193        207        640:  95%|█████████▍| 767/809 [15:02<00:48,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4522     0.1193        281        640:  95%|█████████▌| 769/809 [15:05<00:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4522     0.1194        250        640:  95%|█████████▌| 770/809 [15:06<00:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521     0.1194        195        640:  95%|█████████▌| 772/809 [15:08<00:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521     0.1194        155        640:  96%|█████████▌| 774/809 [15:10<00:40,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117     0.4521     0.1194        178        640:  96%|█████████▌| 775/809 [15:12<00:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.117      0.452     0.1195        188        640:  96%|█████████▌| 777/809 [15:14<00:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118      0.452     0.1196        177        640:  96%|█████████▋| 779/809 [15:16<00:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4519     0.1195        343        640:  97%|█████████▋| 781/809 [15:19<00:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.118     0.4518     0.1196        455        640:  97%|█████████▋| 782/809 [15:20<00:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4517     0.1197        465        640:  97%|█████████▋| 784/809 [15:22<00:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4515     0.1199        500        640:  97%|█████████▋| 786/809 [15:25<00:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.119     0.4514     0.1199        221        640:  97%|█████████▋| 787/809 [15:26<00:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4514     0.1199        161        640:  98%|█████████▊| 789/809 [15:28<00:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G       1.12     0.4512       0.12        338        640:  98%|█████████▊| 791/809 [15:30<00:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4509       0.12        225        640:  98%|█████████▊| 793/809 [15:33<00:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4509       0.12        131        640:  98%|█████████▊| 794/809 [15:34<00:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4509       0.12        242        640:  98%|█████████▊| 796/809 [15:36<00:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4509       0.12        220        640:  99%|█████████▊| 798/809 [15:39<00:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121      0.451     0.1199        180        640:  99%|█████████▉| 799/809 [15:40<00:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4511     0.1199        306        640:  99%|█████████▉| 801/809 [15:42<00:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4509       0.12        494        640:  99%|█████████▉| 803/809 [15:45<00:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4509       0.12        353        640:  99%|█████████▉| 804/809 [15:46<00:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4508       0.12        187        640: 100%|█████████▉| 806/809 [15:48<00:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     76/300      13.6G      1.121     0.4507       0.12        225        640: 100%|█████████▉| 808/809 [15:50<00:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   9%|▊         | 6/69 [00:01<00:14,  4.22it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  20%|██        | 14/69 [00:03<00:13,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  32%|███▏      | 22/69 [00:05<00:11,  4.09it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  43%|████▎     | 30/69 [00:07<00:09,  4.11it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  55%|█████▌    | 38/69 [00:09<00:07,  4.08it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  67%|██████▋   | 46/69 [00:11<00:06,  3.75it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  78%|███████▊  | 54/69 [00:13<00:03,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  90%|████████▉ | 62/69 [00:15<00:01,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:16<00:00,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:16<00:00,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144     0.4764     0.1053        223        640:   0%|          | 1/809 [00:01<15:53,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144     0.4764     0.1053        223        640:   0%|          | 1/809 [00:01<15:53,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.09     0.4844     0.1026        213        640:   0%|          | 2/809 [00:02<15:23,  1.14s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.056     0.4792     0.1045        251        640:   0%|          | 4/809 [00:04<15:58,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.041     0.4695     0.1065        171        640:   1%|          | 6/809 [00:06<15:30,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.074     0.4578     0.1081        361        640:   1%|          | 8/809 [00:09<15:31,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.047     0.4636     0.1047        135        640:   1%|          | 9/809 [00:10<15:45,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.082     0.4574     0.1096        182        640:   1%|▏         | 11/809 [00:12<15:26,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.119     0.4431     0.1085        230        640:   2%|▏         | 13/809 [00:15<15:29,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.139     0.4358     0.1122        407        640:   2%|▏         | 14/809 [00:16<15:30,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.148     0.4337     0.1117        224        640:   2%|▏         | 16/809 [00:18<15:31,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.129     0.4416     0.1117        204        640:   2%|▏         | 18/809 [00:21<15:31,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.115     0.4491     0.1097        131        640:   2%|▏         | 20/809 [00:23<15:23,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.115     0.4486     0.1093        268        640:   3%|▎         | 21/809 [00:24<15:22,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4542     0.1085        161        640:   3%|▎         | 23/809 [00:26<15:18,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.095     0.4543     0.1069        119        640:   3%|▎         | 25/809 [00:29<15:21,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.094     0.4555     0.1065        143        640:   3%|▎         | 26/809 [00:30<15:14,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.102     0.4527     0.1077        233        640:   3%|▎         | 28/809 [00:32<15:09,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.113     0.4489     0.1079         94        640:   4%|▎         | 30/809 [00:35<15:12,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.135     0.4415      0.113        320        640:   4%|▍         | 32/809 [00:37<15:10,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.141     0.4393     0.1141        370        640:   4%|▍         | 33/809 [00:38<15:22,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4434     0.1128        147        640:   4%|▍         | 35/809 [00:40<15:09,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.129     0.4405      0.113        379        640:   5%|▍         | 37/809 [00:43<15:23,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.138     0.4381     0.1143        346        640:   5%|▍         | 38/809 [00:44<15:16,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.131     0.4399     0.1132        301        640:   5%|▍         | 40/809 [00:46<14:50,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.121      0.443      0.114        263        640:   5%|▌         | 42/809 [00:49<15:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.121     0.4424     0.1144        191        640:   5%|▌         | 44/809 [00:51<14:50,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.127     0.4408     0.1171        476        640:   6%|▌         | 45/809 [00:52<14:52,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.137     0.4381     0.1209        166        640:   6%|▌         | 47/809 [00:55<14:47,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.137     0.4376     0.1239        253        640:   6%|▌         | 49/809 [00:57<14:47,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.14     0.4363     0.1245        313        640:   6%|▌         | 50/809 [00:58<14:42,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.138      0.436     0.1241        110        640:   6%|▋         | 52/809 [01:00<14:43,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4392     0.1234        294        640:   7%|▋         | 54/809 [01:03<14:35,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.132     0.4391     0.1232        297        640:   7%|▋         | 56/809 [01:05<14:32,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.128     0.4411     0.1227        147        640:   7%|▋         | 57/809 [01:06<14:47,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.129     0.4411     0.1228        214        640:   7%|▋         | 59/809 [01:09<14:45,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.121     0.4441     0.1217        128        640:   8%|▊         | 61/809 [01:11<14:35,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.12     0.4445     0.1216        332        640:   8%|▊         | 62/809 [01:12<14:30,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4432     0.1258        149        640:   8%|▊         | 64/809 [01:14<14:26,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.128     0.4422     0.1257        275        640:   8%|▊         | 66/809 [01:17<14:25,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.132     0.4411     0.1263        209        640:   8%|▊         | 68/809 [01:19<14:30,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.131     0.4415      0.126        234        640:   9%|▊         | 70/809 [01:21<14:19,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.13     0.4419     0.1258        164        640:   9%|▉         | 71/809 [01:23<14:16,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4408     0.1262        417        640:   9%|▉         | 73/809 [01:25<14:24,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.132     0.4428     0.1262        188        640:   9%|▉         | 75/809 [01:27<14:10,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.132     0.4429     0.1258        281        640:  10%|▉         | 77/809 [01:30<14:05,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.135     0.4423     0.1259        285        640:  10%|▉         | 78/809 [01:31<14:04,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.129     0.4438      0.125        198        640:  10%|▉         | 80/809 [01:33<14:12,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4452      0.124        118        640:  10%|█         | 82/809 [01:36<14:45,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.123     0.4454     0.1237        174        640:  10%|█         | 83/809 [01:37<14:42,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122     0.4461     0.1235        263        640:  11%|█         | 85/809 [01:39<14:33,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4447     0.1233        278        640:  11%|█         | 87/809 [01:42<14:31,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4445     0.1234        246        640:  11%|█         | 88/809 [01:43<14:33,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4444     0.1231        374        640:  11%|█         | 90/809 [01:45<14:54,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122     0.4456     0.1227        181        640:  11%|█         | 91/809 [01:47<14:46,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.119     0.4465     0.1218        294        640:  11%|█▏        | 93/809 [01:49<14:27,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122     0.4459     0.1221        334        640:  12%|█▏        | 95/809 [01:51<14:18,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122     0.4461     0.1216        423        640:  12%|█▏        | 97/809 [01:54<14:07,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4452      0.122        233        640:  12%|█▏        | 98/809 [01:55<13:58,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4456      0.122        197        640:  12%|█▏        | 100/809 [01:57<13:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122     0.4459     0.1214        207        640:  13%|█▎        | 102/809 [01:59<13:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.119     0.4469      0.121         76        640:  13%|█▎        | 104/809 [02:02<13:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.116     0.4479     0.1205        246        640:  13%|█▎        | 105/809 [02:03<13:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.108     0.4495     0.1196        216        640:  13%|█▎        | 107/809 [02:05<13:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.112     0.4485     0.1203        537        640:  13%|█▎        | 109/809 [02:08<13:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.109     0.4493     0.1198        142        640:  14%|█▎        | 110/809 [02:09<13:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.109     0.4494     0.1197        213        640:  14%|█▍        | 112/809 [02:11<13:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.108     0.4493     0.1194        294        640:  14%|█▍        | 114/809 [02:14<13:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.109     0.4489     0.1193        480        640:  14%|█▍        | 116/809 [02:16<13:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.111     0.4482     0.1192        349        640:  14%|█▍        | 117/809 [02:17<13:17,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.118     0.4464     0.1214        391        640:  15%|█▍        | 119/809 [02:19<13:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.118     0.4462     0.1211        349        640:  15%|█▍        | 121/809 [02:22<13:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122      0.445     0.1211        424        640:  15%|█▌        | 122/809 [02:23<13:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4435     0.1208        336        640:  15%|█▌        | 124/809 [02:25<13:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.128     0.4431     0.1203        162        640:  16%|█▌        | 126/809 [02:27<13:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.132     0.4424       0.12        116        640:  16%|█▌        | 128/809 [02:30<13:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.134     0.4417     0.1199        201        640:  16%|█▌        | 129/809 [02:31<13:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.134      0.442     0.1205        160        640:  16%|█▌        | 131/809 [02:34<13:47,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.134     0.4418     0.1203        354        640:  16%|█▋        | 133/809 [02:36<13:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4421     0.1203        199        640:  17%|█▋        | 134/809 [02:37<13:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.134     0.4418     0.1217        208        640:  17%|█▋        | 136/809 [02:39<13:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.132     0.4422     0.1214         64        640:  17%|█▋        | 138/809 [02:42<13:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.135     0.4413     0.1217        323        640:  17%|█▋        | 139/809 [02:43<13:47,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.134     0.4419     0.1214        170        640:  17%|█▋        | 141/809 [02:45<13:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4417     0.1212        285        640:  18%|█▊        | 143/809 [02:48<13:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.136      0.441     0.1212        389        640:  18%|█▊        | 144/809 [02:49<13:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.14     0.4399     0.1219        550        640:  18%|█▊        | 146/809 [02:51<13:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.138     0.4407     0.1215        233        640:  18%|█▊        | 148/809 [02:54<13:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.139     0.4404     0.1215        276        640:  18%|█▊        | 149/809 [02:55<13:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.141     0.4395     0.1212        331        640:  19%|█▊        | 151/809 [02:57<12:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.139     0.4403     0.1208        235        640:  19%|█▉        | 153/809 [03:00<13:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.142     0.4392      0.121        351        640:  19%|█▉        | 155/809 [03:02<12:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.142     0.4389     0.1208        238        640:  19%|█▉        | 156/809 [03:03<12:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.141       0.44     0.1205        176        640:  20%|█▉        | 158/809 [03:05<12:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146      0.439     0.1223        185        640:  20%|█▉        | 160/809 [03:08<12:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.145     0.4396     0.1221        156        640:  20%|█▉        | 161/809 [03:09<12:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146     0.4402     0.1221         97        640:  20%|██        | 163/809 [03:11<13:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146     0.4402     0.1219        328        640:  20%|██        | 165/809 [03:14<12:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.148     0.4399     0.1219        468        640:  21%|██        | 166/809 [03:15<12:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.147     0.4405     0.1216        226        640:  21%|██        | 168/809 [03:17<12:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.148       0.44     0.1217        299        640:  21%|██        | 170/809 [03:20<12:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146      0.441     0.1214        107        640:  21%|██        | 171/809 [03:21<13:00,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.149     0.4401     0.1214        203        640:  21%|██▏       | 173/809 [03:23<12:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.151     0.4398     0.1211        358        640:  22%|██▏       | 175/809 [03:26<13:07,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.15     0.4398     0.1209        228        640:  22%|██▏       | 176/809 [03:27<12:57,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.15       0.44     0.1206        167        640:  22%|██▏       | 178/809 [03:30<12:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.149     0.4404     0.1202         96        640:  22%|██▏       | 180/809 [03:32<12:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.153     0.4393     0.1208        493        640:  22%|██▏       | 181/809 [03:33<12:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.152     0.4397     0.1203        120        640:  23%|██▎       | 183/809 [03:35<12:02,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.153     0.4397       0.12        142        640:  23%|██▎       | 185/809 [03:38<12:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.153     0.4397     0.1201        271        640:  23%|██▎       | 187/809 [03:40<12:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.152     0.4398     0.1199        286        640:  23%|██▎       | 188/809 [03:41<12:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.15     0.4401     0.1195        114        640:  23%|██▎       | 190/809 [03:44<12:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.148     0.4405     0.1193        179        640:  24%|██▎       | 192/809 [03:46<11:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146     0.4411      0.119        192        640:  24%|██▍       | 193/809 [03:47<12:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144     0.4417     0.1189        258        640:  24%|██▍       | 195/809 [03:49<11:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144     0.4418     0.1194        259        640:  24%|██▍       | 197/809 [03:52<11:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144     0.4419     0.1199        110        640:  25%|██▍       | 199/809 [03:54<11:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.143     0.4424     0.1197        255        640:  25%|██▍       | 200/809 [03:55<11:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.138     0.4436     0.1192        159        640:  25%|██▍       | 202/809 [03:58<11:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.136     0.4445     0.1189        165        640:  25%|██▌       | 204/809 [04:00<11:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.135      0.445     0.1188        181        640:  25%|██▌       | 205/809 [04:01<11:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.136      0.445      0.119        240        640:  26%|██▌       | 207/809 [04:03<11:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.135     0.4452      0.119        162        640:  26%|██▌       | 209/809 [04:06<11:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.134     0.4453     0.1192        209        640:  26%|██▌       | 211/809 [04:08<11:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.134     0.4452      0.119        198        640:  26%|██▌       | 212/809 [04:09<11:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.134     0.4452      0.119        388        640:  26%|██▋       | 214/809 [04:11<11:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4454     0.1188        207        640:  27%|██▋       | 216/809 [04:14<11:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4456     0.1188        205        640:  27%|██▋       | 217/809 [04:15<11:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4457     0.1186        389        640:  27%|██▋       | 219/809 [04:17<11:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4455     0.1186        368        640:  27%|██▋       | 221/809 [04:20<11:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.13     0.4459     0.1181        174        640:  28%|██▊       | 223/809 [04:22<11:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.129     0.4462     0.1179        245        640:  28%|██▊       | 224/809 [04:23<11:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.13     0.4458     0.1181        352        640:  28%|██▊       | 226/809 [04:25<11:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.132      0.445     0.1183        310        640:  28%|██▊       | 228/809 [04:28<11:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.136     0.4435     0.1187        268        640:  28%|██▊       | 230/809 [04:30<11:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.136     0.4438     0.1189        127        640:  29%|██▊       | 231/809 [04:31<11:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.138     0.4432     0.1188        196        640:  29%|██▉       | 233/809 [04:34<11:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.138     0.4428     0.1186        184        640:  29%|██▉       | 235/809 [04:36<11:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.141     0.4419     0.1192        584        640:  29%|██▉       | 236/809 [04:37<11:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.142     0.4415     0.1192        287        640:  29%|██▉       | 238/809 [04:40<11:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.142     0.4414     0.1193        431        640:  30%|██▉       | 240/809 [04:42<11:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.143     0.4411     0.1193        193        640:  30%|██▉       | 242/809 [04:44<11:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.142     0.4411     0.1192        334        640:  30%|███       | 244/809 [04:47<10:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.143     0.4408     0.1194        354        640:  30%|███       | 245/809 [04:48<10:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.142     0.4413     0.1194        389        640:  31%|███       | 247/809 [04:50<10:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.145     0.4406     0.1209        237        640:  31%|███       | 249/809 [04:53<11:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146     0.4401     0.1213        396        640:  31%|███       | 250/809 [04:54<11:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146     0.4404     0.1213        123        640:  31%|███       | 252/809 [04:56<10:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.145     0.4408     0.1211        305        640:  31%|███▏      | 254/809 [04:58<10:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.147     0.4402     0.1219        274        640:  32%|███▏      | 255/809 [05:00<10:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146     0.4408     0.1216        253        640:  32%|███▏      | 257/809 [05:02<10:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146     0.4408     0.1215        172        640:  32%|███▏      | 259/809 [05:04<10:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.147     0.4407      0.122        364        640:  32%|███▏      | 261/809 [05:07<10:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.147     0.4404     0.1222        419        640:  32%|███▏      | 262/809 [05:08<10:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.149     0.4401     0.1226        270        640:  33%|███▎      | 264/809 [05:10<10:28,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.148     0.4402     0.1225        182        640:  33%|███▎      | 266/809 [05:13<10:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.148     0.4402     0.1224        167        640:  33%|███▎      | 267/809 [05:14<10:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.151     0.4394     0.1232        171        640:  33%|███▎      | 269/809 [05:16<10:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.154     0.4386     0.1236        623        640:  33%|███▎      | 271/809 [05:18<10:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.154     0.4388     0.1238        372        640:  34%|███▎      | 273/809 [05:21<10:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.155     0.4385     0.1239        442        640:  34%|███▍      | 274/809 [05:22<10:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.158     0.4377     0.1246        195        640:  34%|███▍      | 276/809 [05:24<10:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.163     0.4364     0.1258        438        640:  34%|███▍      | 278/809 [05:27<10:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.164     0.4362     0.1261        339        640:  34%|███▍      | 279/809 [05:28<10:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.167     0.4355     0.1266        316        640:  35%|███▍      | 281/809 [05:30<10:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.167     0.4355     0.1265        227        640:  35%|███▍      | 283/809 [05:33<10:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.168     0.4353     0.1265        306        640:  35%|███▌      | 284/809 [05:34<10:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.168     0.4352     0.1263        393        640:  35%|███▌      | 286/809 [05:36<10:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.167     0.4356     0.1261        182        640:  36%|███▌      | 288/809 [05:39<10:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.166     0.4359      0.126        178        640:  36%|███▌      | 289/809 [05:40<10:36,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.164     0.4366     0.1257        211        640:  36%|███▌      | 291/809 [05:42<10:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.163     0.4368     0.1255        122        640:  36%|███▌      | 292/809 [05:44<10:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.163     0.4369     0.1257        376        640:  36%|███▋      | 294/809 [05:46<10:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.162     0.4369     0.1255        174        640:  37%|███▋      | 296/809 [05:48<10:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.162     0.4369     0.1254        220        640:  37%|███▋      | 298/809 [05:51<10:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.161     0.4373     0.1252        181        640:  37%|███▋      | 299/809 [05:52<09:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.16     0.4375     0.1254        112        640:  37%|███▋      | 301/809 [05:54<10:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.157     0.4382     0.1252        292        640:  37%|███▋      | 303/809 [05:57<09:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.157     0.4381     0.1251        382        640:  38%|███▊      | 304/809 [05:58<09:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.157      0.438      0.125        302        640:  38%|███▊      | 306/809 [06:00<09:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.157     0.4382     0.1249        294        640:  38%|███▊      | 308/809 [06:03<09:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.156     0.4384     0.1247        229        640:  38%|███▊      | 309/809 [06:04<09:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.154     0.4389     0.1244        150        640:  38%|███▊      | 311/809 [06:06<09:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.152     0.4395     0.1242        188        640:  39%|███▊      | 313/809 [06:09<09:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.151     0.4402     0.1239        269        640:  39%|███▉      | 315/809 [06:11<09:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.151     0.4399     0.1239        468        640:  39%|███▉      | 316/809 [06:12<09:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.152     0.4398     0.1239        248        640:  39%|███▉      | 318/809 [06:14<09:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.151     0.4399     0.1237        189        640:  40%|███▉      | 320/809 [06:17<09:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.15     0.4401     0.1235        277        640:  40%|███▉      | 321/809 [06:18<09:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.149     0.4403     0.1233        292        640:  40%|███▉      | 323/809 [06:20<09:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.147      0.441      0.123        190        640:  40%|████      | 325/809 [06:23<09:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146     0.4411     0.1231        362        640:  40%|████      | 327/809 [06:25<09:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144     0.4413     0.1229        258        640:  41%|████      | 328/809 [06:26<09:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144     0.4413     0.1228        260        640:  41%|████      | 330/809 [06:29<09:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146     0.4408     0.1239        532        640:  41%|████      | 331/809 [06:30<09:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144     0.4411     0.1237        131        640:  41%|████      | 333/809 [06:32<09:38,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144     0.4411     0.1236        197        640:  41%|████▏     | 335/809 [06:35<09:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144      0.441     0.1236        306        640:  42%|████▏     | 336/809 [06:36<09:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144      0.441     0.1242        266        640:  42%|████▏     | 338/809 [06:38<09:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.145     0.4408     0.1242        368        640:  42%|████▏     | 340/809 [06:41<09:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.145     0.4406     0.1243        886        640:  42%|████▏     | 341/809 [06:42<09:42,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146     0.4404     0.1246        916        640:  42%|████▏     | 343/809 [06:44<09:33,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.147     0.4404     0.1248        324        640:  43%|████▎     | 345/809 [06:47<09:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.146     0.4406     0.1247        282        640:  43%|████▎     | 347/809 [06:49<09:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.145     0.4408     0.1247        228        640:  43%|████▎     | 348/809 [06:50<08:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.144      0.441     0.1246        206        640:  43%|████▎     | 350/809 [06:53<08:48,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.143     0.4413     0.1243        332        640:  44%|████▎     | 352/809 [06:55<08:45,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.142     0.4414     0.1242        243        640:  44%|████▎     | 353/809 [06:56<08:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.142     0.4415     0.1245        175        640:  44%|████▍     | 355/809 [06:58<08:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.142     0.4417     0.1245        364        640:  44%|████▍     | 357/809 [07:01<08:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.14     0.4419     0.1244        200        640:  44%|████▍     | 359/809 [07:03<08:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.141     0.4417     0.1247        418        640:  44%|████▍     | 360/809 [07:04<08:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.14      0.442     0.1247        491        640:  45%|████▍     | 362/809 [07:07<08:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.139     0.4426     0.1245        274        640:  45%|████▍     | 364/809 [07:09<08:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.139     0.4424     0.1245        330        640:  45%|████▌     | 365/809 [07:10<08:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.139     0.4426     0.1247        159        640:  45%|████▌     | 367/809 [07:12<08:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.14     0.4425     0.1254        308        640:  46%|████▌     | 369/809 [07:15<08:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.14     0.4424     0.1254        487        640:  46%|████▌     | 371/809 [07:17<08:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.14     0.4425     0.1254        292        640:  46%|████▌     | 372/809 [07:18<08:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.139     0.4427     0.1252        359        640:  46%|████▌     | 374/809 [07:20<08:22,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.137     0.4432      0.125        222        640:  46%|████▋     | 376/809 [07:23<08:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.137     0.4434     0.1249        249        640:  47%|████▋     | 377/809 [07:24<08:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.136     0.4436     0.1247        144        640:  47%|████▋     | 379/809 [07:26<08:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.135     0.4438     0.1247        346        640:  47%|████▋     | 381/809 [07:29<08:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.135     0.4438     0.1249        606        640:  47%|████▋     | 383/809 [07:31<08:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.135     0.4437     0.1251        266        640:  47%|████▋     | 384/809 [07:32<08:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.134     0.4439     0.1249        169        640:  48%|████▊     | 386/809 [07:35<08:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4442     0.1248        196        640:  48%|████▊     | 388/809 [07:37<08:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4443     0.1247        290        640:  48%|████▊     | 390/809 [07:39<08:02,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4445     0.1246        137        640:  48%|████▊     | 391/809 [07:40<08:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4443     0.1246        380        640:  49%|████▊     | 393/809 [07:43<08:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.134     0.4442     0.1247        464        640:  49%|████▉     | 395/809 [07:45<08:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4443     0.1247        502        640:  49%|████▉     | 397/809 [07:47<07:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.135     0.4438     0.1251        481        640:  49%|████▉     | 399/809 [07:50<08:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.135     0.4438      0.125        211        640:  49%|████▉     | 400/809 [07:51<08:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.134     0.4441     0.1248         97        640:  50%|████▉     | 402/809 [07:53<08:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4441     0.1247        178        640:  50%|████▉     | 404/809 [07:56<08:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133      0.444     0.1246        211        640:  50%|█████     | 405/809 [07:57<08:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.133     0.4443     0.1245        233        640:  50%|█████     | 407/809 [07:59<07:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.131     0.4447     0.1243        141        640:  51%|█████     | 409/809 [08:02<07:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.131     0.4447     0.1243        456        640:  51%|█████     | 410/809 [08:03<07:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.13     0.4449      0.124        213        640:  51%|█████     | 412/809 [08:05<07:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.131     0.4448     0.1243        544        640:  51%|█████     | 414/809 [08:07<07:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.13     0.4451      0.124        339        640:  51%|█████▏    | 416/809 [08:10<07:31,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.129     0.4452     0.1239        224        640:  52%|█████▏    | 417/809 [08:11<07:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.128     0.4457     0.1237        193        640:  52%|█████▏    | 419/809 [08:13<07:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.127     0.4459     0.1235        128        640:  52%|█████▏    | 421/809 [08:15<07:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4465     0.1232        119        640:  52%|█████▏    | 423/809 [08:18<07:22,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4463     0.1235        714        640:  52%|█████▏    | 424/809 [08:19<07:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4466     0.1233        250        640:  53%|█████▎    | 426/809 [08:21<07:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.123     0.4471     0.1231        208        640:  53%|█████▎    | 428/809 [08:24<07:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122     0.4475     0.1229         80        640:  53%|█████▎    | 429/809 [08:25<07:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.121     0.4478     0.1228        270        640:  53%|█████▎    | 431/809 [08:27<07:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122     0.4475     0.1227        243        640:  54%|█████▎    | 433/809 [08:30<07:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122     0.4473     0.1227        357        640:  54%|█████▎    | 434/809 [08:31<07:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.123     0.4472     0.1229        480        640:  54%|█████▍    | 436/809 [08:33<07:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4468     0.1233        331        640:  54%|█████▍    | 438/809 [08:35<07:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4469     0.1231        160        640:  54%|█████▍    | 440/809 [08:38<07:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4468      0.123        313        640:  55%|█████▍    | 441/809 [08:39<07:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122     0.4472     0.1228        238        640:  55%|█████▍    | 443/809 [08:41<07:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122     0.4473     0.1228        208        640:  55%|█████▌    | 445/809 [08:44<07:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.123      0.447      0.123        428        640:  55%|█████▌    | 446/809 [08:45<07:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.123      0.447      0.123        207        640:  55%|█████▌    | 448/809 [08:47<06:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4469     0.1231        250        640:  56%|█████▌    | 450/809 [08:49<06:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4466     0.1231        310        640:  56%|█████▌    | 452/809 [08:52<06:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4467     0.1231        317        640:  56%|█████▌    | 453/809 [08:53<06:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4464     0.1232        185        640:  56%|█████▌    | 455/809 [08:55<06:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4464      0.123        386        640:  56%|█████▋    | 457/809 [08:57<06:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4464     0.1228        352        640:  57%|█████▋    | 459/809 [09:00<06:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4464     0.1227        278        640:  57%|█████▋    | 460/809 [09:01<06:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4465     0.1226        268        640:  57%|█████▋    | 462/809 [09:03<06:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122     0.4471     0.1223        121        640:  57%|█████▋    | 464/809 [09:06<06:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.123     0.4471     0.1224        365        640:  57%|█████▋    | 465/809 [09:07<06:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.123      0.447     0.1223        199        640:  58%|█████▊    | 467/809 [09:09<06:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4465     0.1222        150        640:  58%|█████▊    | 469/809 [09:11<06:31,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4462     0.1222        242        640:  58%|█████▊    | 471/809 [09:14<06:28,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4463     0.1221        144        640:  58%|█████▊    | 472/809 [09:15<06:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4461     0.1222        152        640:  59%|█████▊    | 474/809 [09:17<06:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4464      0.122        169        640:  59%|█████▉    | 476/809 [09:20<06:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4464      0.122        405        640:  59%|█████▉    | 477/809 [09:21<06:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4466     0.1218        162        640:  59%|█████▉    | 479/809 [09:23<06:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.123     0.4471     0.1216        297        640:  59%|█████▉    | 481/809 [09:26<06:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4468     0.1216        310        640:  60%|█████▉    | 482/809 [09:27<06:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.125     0.4465     0.1215        230        640:  60%|█████▉    | 484/809 [09:29<06:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4462     0.1215        115        640:  60%|██████    | 486/809 [09:32<06:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4459     0.1214        214        640:  60%|██████    | 488/809 [09:34<06:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126     0.4459     0.1214        192        640:  60%|██████    | 489/809 [09:35<06:33,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.126      0.446     0.1212        185        640:  61%|██████    | 491/809 [09:38<06:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4464     0.1212        158        640:  61%|██████    | 493/809 [09:40<06:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.124     0.4465     0.1211        300        640:  61%|██████    | 494/809 [09:41<06:25,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.123     0.4467      0.121        224        640:  61%|██████▏   | 496/809 [09:44<06:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122      0.447     0.1209        143        640:  62%|██████▏   | 498/809 [09:46<06:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122      0.447     0.1209        240        640:  62%|██████▏   | 499/809 [09:47<06:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.122      0.447     0.1208        335        640:  62%|██████▏   | 501/809 [09:50<06:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.121     0.4474     0.1206        111        640:  62%|██████▏   | 503/809 [09:52<05:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.12     0.4476     0.1205        188        640:  62%|██████▏   | 504/809 [09:53<05:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.12     0.4476     0.1206        248        640:  63%|██████▎   | 506/809 [09:55<05:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.12     0.4477     0.1207        282        640:  63%|██████▎   | 508/809 [09:58<05:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.12     0.4478     0.1207        254        640:  63%|██████▎   | 510/809 [10:00<05:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.12     0.4478     0.1206        224        640:  63%|██████▎   | 511/809 [10:01<05:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.118      0.448     0.1205         82        640:  63%|██████▎   | 513/809 [10:04<05:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.117     0.4486     0.1204        248        640:  64%|██████▎   | 515/809 [10:06<05:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.117     0.4486     0.1204        291        640:  64%|██████▍   | 516/809 [10:07<05:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.116     0.4489     0.1202        305        640:  64%|██████▍   | 518/809 [10:10<05:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.116     0.4487     0.1201        285        640:  64%|██████▍   | 520/809 [10:12<05:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.116     0.4487     0.1201        341        640:  64%|██████▍   | 521/809 [10:13<05:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.116     0.4486     0.1201        382        640:  65%|██████▍   | 523/809 [10:15<05:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.116     0.4488     0.1199        132        640:  65%|██████▍   | 525/809 [10:18<05:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.114     0.4492     0.1197         96        640:  65%|██████▌   | 527/809 [10:20<05:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.114     0.4493     0.1197        278        640:  65%|██████▌   | 528/809 [10:21<05:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.114     0.4493     0.1197        344        640:  66%|██████▌   | 530/809 [10:24<05:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.114     0.4493     0.1196        302        640:  66%|██████▌   | 532/809 [10:26<05:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.115     0.4491     0.1196        351        640:  66%|██████▌   | 533/809 [10:27<05:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.115     0.4491     0.1195        115        640:  66%|██████▌   | 535/809 [10:29<05:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.117     0.4485     0.1196        455        640:  66%|██████▋   | 537/809 [10:32<05:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.117     0.4486     0.1195        218        640:  67%|██████▋   | 539/809 [10:34<05:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.117     0.4485     0.1197        315        640:  67%|██████▋   | 541/809 [10:36<05:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.117     0.4486     0.1195        262        640:  67%|██████▋   | 543/809 [10:39<05:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.116     0.4487     0.1195        204        640:  67%|██████▋   | 544/809 [10:40<05:04,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.115      0.449     0.1193        187        640:  67%|██████▋   | 546/809 [10:42<05:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.115      0.449     0.1193        198        640:  68%|██████▊   | 548/809 [10:44<05:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.115     0.4493     0.1192        172        640:  68%|██████▊   | 549/809 [10:46<05:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.114     0.4494     0.1191        426        640:  68%|██████▊   | 551/809 [10:48<05:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.113     0.4498      0.119         91        640:  68%|██████▊   | 553/809 [10:50<05:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.112     0.4499     0.1189        205        640:  69%|██████▊   | 555/809 [10:53<04:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.112     0.4501     0.1188        202        640:  69%|██████▊   | 556/809 [10:54<05:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.11     0.4506     0.1186        133        640:  69%|██████▉   | 558/809 [10:56<04:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.108      0.451     0.1184        182        640:  69%|██████▉   | 560/809 [10:59<04:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.108     0.4511     0.1183        144        640:  69%|██████▉   | 561/809 [11:00<04:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4512     0.1183        272        640:  70%|██████▉   | 563/809 [11:02<04:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4513     0.1183        286        640:  70%|██████▉   | 565/809 [11:05<04:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.108     0.4513     0.1183        279        640:  70%|██████▉   | 566/809 [11:06<04:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4518     0.1182        150        640:  70%|███████   | 568/809 [11:08<04:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4516     0.1181        408        640:  70%|███████   | 570/809 [11:10<04:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4517      0.118        228        640:  71%|███████   | 572/809 [11:13<04:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4516      0.118        126        640:  71%|███████   | 573/809 [11:14<04:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4517     0.1179        268        640:  71%|███████   | 575/809 [11:16<04:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4517     0.1178        238        640:  71%|███████▏  | 577/809 [11:19<04:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4518     0.1177        227        640:  71%|███████▏  | 578/809 [11:20<04:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4518      0.118        160        640:  72%|███████▏  | 580/809 [11:22<04:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.108     0.4517      0.118        275        640:  72%|███████▏  | 582/809 [11:25<04:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4518      0.118        176        640:  72%|███████▏  | 583/809 [11:26<04:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4521     0.1178        140        640:  72%|███████▏  | 585/809 [11:28<04:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4521     0.1177        328        640:  73%|███████▎  | 587/809 [11:31<04:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4521     0.1178        262        640:  73%|███████▎  | 589/809 [11:33<04:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4522     0.1178        167        640:  73%|███████▎  | 590/809 [11:34<04:11,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4522     0.1177        161        640:  73%|███████▎  | 592/809 [11:36<04:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4516      0.118        257        640:  73%|███████▎  | 593/809 [11:39<04:30,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4516      0.118        414        640:  74%|███████▎  | 595/809 [11:40<04:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.107     0.4518     0.1179        235        640:  74%|███████▍  | 597/809 [11:42<04:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4519      0.118        183        640:  74%|███████▍  | 598/809 [11:44<04:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4518     0.1181         59        640:  74%|███████▍  | 600/809 [11:46<04:16,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106      0.452      0.118         82        640:  74%|███████▍  | 602/809 [11:49<04:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4521     0.1181         77        640:  75%|███████▍  | 604/809 [11:51<04:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4521      0.118        207        640:  75%|███████▍  | 605/809 [11:52<03:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.104     0.4521      0.118        166        640:  75%|███████▌  | 607/809 [11:54<03:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4519     0.1181        173        640:  75%|███████▌  | 609/809 [11:57<03:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4518     0.1181        390        640:  76%|███████▌  | 611/809 [11:59<03:47,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4517      0.118        205        640:  76%|███████▌  | 612/809 [12:00<03:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4519     0.1179        263        640:  76%|███████▌  | 614/809 [12:02<03:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4519     0.1178        248        640:  76%|███████▌  | 616/809 [12:05<03:41,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4521     0.1177         85        640:  76%|███████▋  | 617/809 [12:06<03:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4518     0.1177        309        640:  77%|███████▋  | 619/809 [12:08<03:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4521     0.1176        210        640:  77%|███████▋  | 621/809 [12:10<03:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4519     0.1175        281        640:  77%|███████▋  | 623/809 [12:13<03:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4519     0.1175        188        640:  77%|███████▋  | 624/809 [12:14<03:31,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4518     0.1174        316        640:  77%|███████▋  | 626/809 [12:16<03:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4521     0.1173        201        640:  78%|███████▊  | 628/809 [12:19<03:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105      0.452     0.1174        503        640:  78%|███████▊  | 629/809 [12:20<03:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105      0.452     0.1172        160        640:  78%|███████▊  | 631/809 [12:22<03:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4521     0.1173        381        640:  78%|███████▊  | 633/809 [12:25<03:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105      0.452     0.1173        152        640:  78%|███████▊  | 635/809 [12:27<03:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4523     0.1173        110        640:  79%|███████▊  | 636/809 [12:28<03:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4522     0.1172        239        640:  79%|███████▉  | 638/809 [12:31<03:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4519     0.1172        159        640:  79%|███████▉  | 640/809 [12:33<03:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4518     0.1172        239        640:  79%|███████▉  | 641/809 [12:34<03:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4519     0.1171        276        640:  79%|███████▉  | 643/809 [12:36<03:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106      0.452      0.117        234        640:  80%|███████▉  | 645/809 [12:39<03:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106      0.452     0.1169        298        640:  80%|███████▉  | 647/809 [12:41<03:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4521     0.1169        247        640:  80%|████████  | 648/809 [12:42<03:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4521     0.1168        306        640:  80%|████████  | 650/809 [12:44<03:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4522     0.1168        266        640:  81%|████████  | 652/809 [12:47<03:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4523     0.1167        161        640:  81%|████████  | 653/809 [12:48<03:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4525     0.1167        121        640:  81%|████████  | 655/809 [12:50<03:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.106     0.4523     0.1168        134        640:  81%|████████  | 657/809 [12:53<03:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4525     0.1167        167        640:  81%|████████▏ | 658/809 [12:54<03:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4526     0.1166        242        640:  82%|████████▏ | 660/809 [12:56<02:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.104     0.4527     0.1166        410        640:  82%|████████▏ | 662/809 [12:59<02:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4526     0.1167        309        640:  82%|████████▏ | 664/809 [13:01<02:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4527     0.1167        222        640:  82%|████████▏ | 665/809 [13:02<02:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.104     0.4528     0.1166        309        640:  82%|████████▏ | 667/809 [13:05<02:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.104     0.4529     0.1166        167        640:  83%|████████▎ | 669/809 [13:07<02:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.104     0.4527     0.1167        215        640:  83%|████████▎ | 670/809 [13:08<02:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.104     0.4528     0.1166        179        640:  83%|████████▎ | 672/809 [13:10<02:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.104     0.4527     0.1165        385        640:  83%|████████▎ | 674/809 [13:13<02:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.104     0.4526     0.1165        314        640:  84%|████████▎ | 676/809 [13:15<02:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.104     0.4526     0.1165        336        640:  84%|████████▎ | 677/809 [13:16<02:30,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.105     0.4525     0.1165        529        640:  84%|████████▍ | 679/809 [13:19<02:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.104     0.4527     0.1165        200        640:  84%|████████▍ | 681/809 [13:21<02:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.103     0.4528     0.1164        224        640:  84%|████████▍ | 682/809 [13:22<02:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.103      0.453     0.1164        143        640:  85%|████████▍ | 684/809 [13:24<02:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.102     0.4532     0.1163         57        640:  85%|████████▍ | 686/809 [13:27<02:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4534     0.1162        111        640:  85%|████████▌ | 688/809 [13:29<02:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G        1.1     0.4536     0.1161        177        640:  85%|████████▌ | 689/809 [13:30<02:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4535     0.1161        146        640:  85%|████████▌ | 691/809 [13:33<02:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4535      0.116        144        640:  86%|████████▌ | 693/809 [13:35<02:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4534     0.1159        349        640:  86%|████████▌ | 695/809 [13:37<02:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4533     0.1158        171        640:  86%|████████▌ | 697/809 [13:40<02:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4532     0.1159        368        640:  86%|████████▋ | 698/809 [13:41<02:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.102     0.4531      0.116        232        640:  87%|████████▋ | 700/809 [13:43<02:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4532     0.1159        196        640:  87%|████████▋ | 702/809 [13:45<02:02,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4533     0.1158        146        640:  87%|████████▋ | 704/809 [13:48<02:00,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4533     0.1158        123        640:  87%|████████▋ | 705/809 [13:49<02:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.102     0.4529     0.1161        415        640:  87%|████████▋ | 707/809 [13:51<02:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.102     0.4529     0.1161        288        640:  88%|████████▊ | 709/809 [13:54<01:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.102     0.4528     0.1161        427        640:  88%|████████▊ | 710/809 [13:55<01:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.103     0.4527     0.1161        280        640:  88%|████████▊ | 712/809 [13:57<01:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.102     0.4529     0.1161        271        640:  88%|████████▊ | 714/809 [13:59<01:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.102     0.4529     0.1161        337        640:  88%|████████▊ | 715/809 [14:01<01:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.102     0.4529     0.1161        437        640:  89%|████████▊ | 717/809 [14:03<01:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4532      0.116        112        640:  89%|████████▉ | 719/809 [14:05<01:43,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4532     0.1159        226        640:  89%|████████▉ | 721/809 [14:08<01:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4532     0.1159        231        640:  89%|████████▉ | 722/809 [14:09<01:40,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.101     0.4535     0.1158        213        640:  89%|████████▉ | 724/809 [14:11<01:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.099     0.4537     0.1157        187        640:  90%|████████▉ | 726/809 [14:14<01:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.099     0.4539     0.1157        245        640:  90%|████████▉ | 728/809 [14:16<01:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.098     0.4539     0.1157        140        640:  90%|█████████ | 729/809 [14:17<01:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.098      0.454     0.1156        186        640:  90%|█████████ | 731/809 [14:19<01:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.098      0.454     0.1156        304        640:  91%|█████████ | 733/809 [14:22<01:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.098     0.4539     0.1156        475        640:  91%|█████████ | 734/809 [14:23<01:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.098     0.4538     0.1155        269        640:  91%|█████████ | 736/809 [14:25<01:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.097     0.4539     0.1154        126        640:  91%|█████████ | 738/809 [14:28<01:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.097      0.454     0.1153        138        640:  91%|█████████▏| 740/809 [14:30<01:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.097      0.454     0.1153        223        640:  92%|█████████▏| 741/809 [14:31<01:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.097     0.4539     0.1157        493        640:  92%|█████████▏| 743/809 [14:33<01:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.097      0.454     0.1156        296        640:  92%|█████████▏| 745/809 [14:36<01:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.097     0.4541     0.1157        314        640:  92%|█████████▏| 746/809 [14:37<01:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.096     0.4542     0.1155        259        640:  92%|█████████▏| 748/809 [14:39<01:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.097     0.4541     0.1157        290        640:  93%|█████████▎| 750/809 [14:42<01:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.096     0.4542     0.1156        214        640:  93%|█████████▎| 751/809 [14:43<01:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.098     0.4539     0.1162        650        640:  93%|█████████▎| 753/809 [14:45<01:09,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.098     0.4537     0.1163        474        640:  93%|█████████▎| 755/809 [14:48<01:08,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.098     0.4538     0.1163        283        640:  93%|█████████▎| 756/809 [14:49<01:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.098     0.4538     0.1163        243        640:  94%|█████████▎| 758/809 [14:51<01:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.097      0.454     0.1161        205        640:  94%|█████████▍| 760/809 [14:54<00:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.097      0.454     0.1161        204        640:  94%|█████████▍| 761/809 [14:55<00:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.097     0.4541      0.116        236        640:  94%|█████████▍| 763/809 [14:57<00:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.096     0.4545     0.1159        123        640:  95%|█████████▍| 765/809 [15:00<00:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.095     0.4547     0.1157        153        640:  95%|█████████▍| 767/809 [15:02<00:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.094     0.4548     0.1157        254        640:  95%|█████████▍| 768/809 [15:03<00:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.093     0.4551     0.1156        202        640:  95%|█████████▌| 770/809 [15:05<00:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.092     0.4554     0.1155        161        640:  95%|█████████▌| 772/809 [15:08<00:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.092     0.4555     0.1154        113        640:  96%|█████████▌| 773/809 [15:09<00:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.091     0.4556     0.1153        289        640:  96%|█████████▌| 775/809 [15:11<00:41,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.091     0.4559     0.1152        309        640:  96%|█████████▌| 777/809 [15:14<00:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.091     0.4559     0.1151        249        640:  96%|█████████▌| 778/809 [15:15<00:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.09     0.4561     0.1151        256        640:  96%|█████████▋| 780/809 [15:17<00:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.09     0.4562      0.115        291        640:  97%|█████████▋| 782/809 [15:20<00:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.09     0.4562      0.115        328        640:  97%|█████████▋| 784/809 [15:22<00:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.09     0.4562      0.115        213        640:  97%|█████████▋| 785/809 [15:23<00:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.09     0.4562      0.115        401        640:  97%|█████████▋| 787/809 [15:26<00:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.09     0.4564     0.1149        182        640:  98%|█████████▊| 789/809 [15:28<00:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.09     0.4564     0.1149        283        640:  98%|█████████▊| 790/809 [15:29<00:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.089     0.4565      0.115        170        640:  98%|█████████▊| 792/809 [15:31<00:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.089     0.4566     0.1149        155        640:  98%|█████████▊| 794/809 [15:34<00:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.09     0.4565      0.115        229        640:  98%|█████████▊| 796/809 [15:36<00:14,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G       1.09     0.4564      0.115        331        640:  99%|█████████▊| 797/809 [15:37<00:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.092     0.4559     0.1152        401        640:  99%|█████████▉| 799/809 [15:39<00:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.093     0.4556     0.1153        454        640:  99%|█████████▉| 801/809 [15:42<00:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.093     0.4556     0.1153        173        640:  99%|█████████▉| 802/809 [15:43<00:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.092     0.4558     0.1152        213        640:  99%|█████████▉| 804/809 [15:45<00:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.092     0.4558     0.1151        309        640: 100%|█████████▉| 806/809 [15:48<00:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     77/300      13.7G      1.092     0.4557     0.1151        322        640: 100%|█████████▉| 808/809 [15:50<00:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   6%|▌         | 4/69 [00:00<00:15,  4.23it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  17%|█▋        | 12/69 [00:02<00:13,  4.11it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  29%|██▉       | 20/69 [00:04<00:11,  4.10it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  41%|████      | 28/69 [00:06<00:10,  3.79it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  52%|█████▏    | 36/69 [00:09<00:08,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  64%|██��███▍   | 44/69 [00:11<00:06,  3.92it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  74%|███████▍  | 51/69 [00:12<00:04,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  87%|████████▋ | 60/69 [00:15<00:02,  4.13it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  99%|█████████▊| 68/69 [00:17<00:00,  4.14it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  99%|█████████▊| 68/69 [00:17<00:00,  4.14it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.4G      1.354     0.3817     0.1015        296        640:   0%|          | 2/809 [00:02<15:40,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.4G      1.307     0.3969    0.09975        231        640:   0%|          | 4/809 [00:04<15:26,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.4G      1.373      0.374     0.1066        429        640:   1%|          | 5/809 [00:05<15:27,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.4G      1.371     0.3817     0.1193        136        640:   1%|          | 7/809 [00:08<15:22,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.343     0.3829     0.1172        200        640:   1%|          | 9/809 [00:10<15:48,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.313     0.3902     0.1148        222        640:   1%|          | 10/809 [00:11<15:34,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.242     0.4071     0.1081        426        640:   1%|▏         | 12/809 [00:14<15:32,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.183     0.4299     0.1047        173        640:   2%|▏         | 14/809 [00:16<15:30,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.163     0.4382     0.1028        176        640:   2%|▏         | 15/809 [00:17<15:32,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.216     0.4302     0.1259        460        640:   2%|▏         | 17/809 [00:20<16:30,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.175     0.4418     0.1212        217        640:   2%|▏         | 19/809 [00:22<15:56,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.156     0.4461     0.1184        142        640:   2%|▏         | 20/809 [00:23<15:44,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.137     0.4482     0.1152        321        640:   3%|▎         | 22/809 [00:26<15:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.104     0.4545     0.1114         70        640:   3%|▎         | 24/809 [00:28<15:42,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4549     0.1099        144        640:   3%|▎         | 25/809 [00:29<15:37,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4501     0.1104        371        640:   3%|▎         | 27/809 [00:31<15:15,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.143     0.4411     0.1176        508        640:   4%|▎         | 29/809 [00:34<15:17,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.157     0.4375     0.1177        355        640:   4%|▍         | 31/809 [00:36<15:13,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.168      0.434     0.1198        465        640:   4%|▍         | 32/809 [00:37<15:14,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.162     0.4387     0.1178        233        640:   4%|▍         | 34/809 [00:40<14:54,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.159      0.439     0.1163        285        640:   4%|▍         | 36/809 [00:42<14:47,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.166     0.4381     0.1195        428        640:   5%|▍         | 38/809 [00:44<14:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.156     0.4414     0.1181        189        640:   5%|▍         | 40/809 [00:47<14:56,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.161     0.4397     0.1194        201        640:   5%|▌         | 41/809 [00:48<15:03,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.162     0.4387     0.1188        250        640:   5%|▌         | 43/809 [00:50<15:06,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.161     0.4395     0.1174        180        640:   6%|▌         | 45/809 [00:53<14:54,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.158     0.4406     0.1165        163        640:   6%|▌         | 47/809 [00:55<14:40,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.149      0.443     0.1152         89        640:   6%|▌         | 48/809 [00:56<14:52,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.139     0.4448     0.1142        336        640:   6%|▌         | 50/809 [00:58<14:40,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.133      0.448     0.1139        166        640:   6%|▋         | 52/809 [01:01<14:36,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.132     0.4481     0.1136        450        640:   7%|▋         | 53/809 [01:02<14:28,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.121     0.4488     0.1127        137        640:   7%|▋         | 55/809 [01:04<14:35,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4514     0.1126        412        640:   7%|▋         | 57/809 [01:06<14:38,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4528     0.1122        171        640:   7%|▋         | 59/809 [01:09<14:35,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4522     0.1124        341        640:   7%|▋         | 60/809 [01:10<14:28,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.103      0.453     0.1115        198        640:   8%|▊         | 62/809 [01:12<14:31,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4509     0.1125        383        640:   8%|▊         | 64/809 [01:15<14:37,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.118     0.4484     0.1134        248        640:   8%|▊         | 66/809 [01:17<14:14,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.118     0.4479     0.1131        280        640:   8%|▊         | 67/809 [01:18<14:16,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.123     0.4469     0.1126        170        640:   9%|▊         | 69/809 [01:20<14:12,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.127      0.446     0.1127        333        640:   9%|▉         | 71/809 [01:23<14:28,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.133     0.4444     0.1131        485        640:   9%|▉         | 72/809 [01:24<14:40,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.13     0.4452     0.1129        107        640:   9%|▉         | 74/809 [01:26<14:20,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.126     0.4463     0.1121        291        640:   9%|▉         | 76/809 [01:29<14:39,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.124     0.4473     0.1118        151        640:  10%|▉         | 77/809 [01:30<14:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.118     0.4498     0.1114        306        640:  10%|▉         | 79/809 [01:32<14:34,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.118     0.4494     0.1111        302        640:  10%|█         | 81/809 [01:35<14:25,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.12     0.4491      0.111        289        640:  10%|█         | 83/809 [01:37<14:19,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.121     0.4484     0.1108        174        640:  10%|█         | 84/809 [01:38<14:38,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.123     0.4477     0.1113        485        640:  11%|█         | 86/809 [01:41<14:48,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.124      0.447     0.1112        423        640:  11%|█         | 87/809 [01:42<14:36,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.122     0.4468     0.1107        291        640:  11%|█         | 89/809 [01:44<14:24,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4479     0.1103        205        640:  11%|█         | 91/809 [01:47<13:57,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112     0.4482     0.1097        241        640:  11%|█▏        | 92/809 [01:48<13:52,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4493     0.1092        327        640:  12%|█▏        | 94/809 [01:50<14:20,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.106     0.4507     0.1094        141        640:  12%|█▏        | 96/809 [01:53<14:04,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4494       0.11        463        640:  12%|█▏        | 98/809 [01:55<13:57,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4497     0.1097        228        640:  12%|█▏        | 99/809 [01:56<13:55,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4509     0.1093        122        640:  12%|█▏        | 101/809 [01:58<13:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.119     0.4472     0.1133        299        640:  13%|█▎        | 103/809 [02:01<13:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.12     0.4472     0.1132        133        640:  13%|█▎        | 104/809 [02:02<14:17,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4487     0.1125        175        640:  13%|█▎        | 106/809 [02:04<13:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4491     0.1122        218        640:  13%|█▎        | 108/809 [02:07<13:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4501     0.1117        137        640:  14%|█▎        | 110/809 [02:09<13:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.103     0.4506     0.1115        131        640:  14%|█▎        | 111/809 [02:10<13:25,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.102     0.4505     0.1112        182        640:  14%|█▍        | 113/809 [02:13<13:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.104     0.4502     0.1117        409        640:  14%|█▍        | 115/809 [02:15<13:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101      0.451     0.1113        180        640:  14%|█▍        | 116/809 [02:16<13:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101     0.4507     0.1109        276        640:  15%|█▍        | 118/809 [02:18<13:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.102     0.4505      0.111        284        640:  15%|█▍        | 120/809 [02:21<13:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101     0.4499     0.1109        121        640:  15%|█▌        | 122/809 [02:23<13:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4501     0.1107        294        640:  15%|█▌        | 123/809 [02:24<13:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.095     0.4505     0.1102        268        640:  15%|█▌        | 125/809 [02:27<13:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.093     0.4509     0.1099        198        640:  16%|█▌        | 127/809 [02:29<13:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.092     0.4513     0.1098        282        640:  16%|█▌        | 128/809 [02:30<13:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.089     0.4516     0.1094        158        640:  16%|█▌        | 130/809 [02:33<13:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.087     0.4524     0.1091        294        640:  16%|█▋        | 132/809 [02:35<13:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.086     0.4526      0.109        276        640:  17%|█▋        | 134/809 [02:37<13:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.085     0.4532     0.1088        260        640:  17%|█▋        | 135/809 [02:38<12:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.086     0.4532     0.1088        178        640:  17%|█▋        | 137/809 [02:41<13:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.088     0.4532     0.1085        300        640:  17%|█▋        | 139/809 [02:43<12:52,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.089     0.4529     0.1085        299        640:  17%|█▋        | 140/809 [02:44<12:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.092     0.4523     0.1096        285        640:  18%|█▊        | 142/809 [02:46<12:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.096     0.4516     0.1101        178        640:  18%|█▊        | 144/809 [02:49<13:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.094     0.4529     0.1099        238        640:  18%|█▊        | 146/809 [02:51<13:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.095     0.4528     0.1102        179        640:  18%|█▊        | 147/809 [02:52<12:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.092     0.4537     0.1099        199        640:  18%|█▊        | 149/809 [02:55<12:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.091     0.4543     0.1096        281        640:  19%|█▊        | 151/809 [02:57<12:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.09     0.4545     0.1096        182        640:  19%|█▉        | 152/809 [02:58<13:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.09     0.4547     0.1094        288        640:  19%|█▉        | 154/809 [03:01<12:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.091     0.4548     0.1093        305        640:  19%|█▉        | 156/809 [03:03<12:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.095     0.4539     0.1095        430        640:  20%|█▉        | 158/809 [03:05<12:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.093     0.4542     0.1093        110        640:  20%|█▉        | 159/809 [03:06<12:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.094     0.4538     0.1093        273        640:  20%|█▉        | 161/809 [03:09<12:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.094     0.4543     0.1093        258        640:  20%|██        | 163/809 [03:11<12:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.095     0.4538     0.1093        291        640:  20%|██        | 164/809 [03:12<12:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.094     0.4536     0.1091        277        640:  21%|██        | 166/809 [03:15<12:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.091     0.4546     0.1088        132        640:  21%|██        | 168/809 [03:17<12:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.088     0.4549     0.1085        127        640:  21%|██        | 169/809 [03:18<12:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.087     0.4554     0.1082        208        640:  21%|██        | 171/809 [03:21<12:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.089     0.4548     0.1088        283        640:  21%|██▏       | 173/809 [03:23<12:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.09     0.4542     0.1088        375        640:  22%|██▏       | 175/809 [03:25<12:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.093     0.4534     0.1094        474        640:  22%|██▏       | 176/809 [03:27<12:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.096     0.4524     0.1094        417        640:  22%|██▏       | 178/809 [03:29<12:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.097     0.4521     0.1092        165        640:  22%|██▏       | 180/809 [03:31<12:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.095     0.4525     0.1092        227        640:  22%|██▏       | 181/809 [03:32<12:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4513     0.1108        405        640:  23%|██▎       | 183/809 [03:35<12:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.102     0.4506     0.1122        240        640:  23%|██▎       | 185/809 [03:37<12:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101     0.4504      0.112        306        640:  23%|██▎       | 186/809 [03:38<12:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4518     0.1118        160        640:  23%|██▎       | 188/809 [03:41<12:44,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4514     0.1119        395        640:  23%|██▎       | 189/809 [03:42<12:38,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4512      0.112        280        640:  24%|██▎       | 191/809 [03:45<12:45,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.102     0.4505      0.113        386        640:  24%|██▍       | 193/809 [03:47<12:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.103     0.4503     0.1131        355        640:  24%|██▍       | 194/809 [03:48<12:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.103     0.4504     0.1131        344        640:  24%|██▍       | 196/809 [03:51<12:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4499     0.1132        328        640:  24%|██▍       | 198/809 [03:53<12:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.106     0.4497     0.1134        229        640:  25%|██▍       | 199/809 [03:54<12:25,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4498     0.1137        338        640:  25%|██▍       | 201/809 [03:57<12:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4496     0.1134        334        640:  25%|██▌       | 203/809 [03:59<11:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.104     0.4501     0.1134        225        640:  25%|██▌       | 205/809 [04:01<11:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4498     0.1134        387        640:  26%|██▌       | 207/809 [04:04<11:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.103     0.4505     0.1131        101        640:  26%|██▌       | 208/809 [04:05<11:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4499     0.1139        270        640:  26%|██▌       | 210/809 [04:07<11:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.106     0.4499      0.114        190        640:  26%|██▌       | 212/809 [04:10<11:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4501     0.1138        209        640:  26%|██▋       | 214/809 [04:12<11:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4501     0.1137        124        640:  27%|██▋       | 215/809 [04:13<11:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4503     0.1137        202        640:  27%|██▋       | 217/809 [04:15<11:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.104     0.4514     0.1133        183        640:  27%|██▋       | 219/809 [04:18<11:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101     0.4521     0.1131        149        640:  27%|██▋       | 220/809 [04:19<11:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4529     0.1128        180        640:  27%|██▋       | 222/809 [04:21<11:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.097     0.4533     0.1126        188        640:  28%|██▊       | 224/809 [04:24<11:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4531     0.1131        157        640:  28%|██▊       | 226/809 [04:26<11:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4531      0.113        139        640:  28%|██▊       | 227/809 [04:27<11:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099      0.453     0.1129        264        640:  28%|██▊       | 229/809 [04:29<11:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4525     0.1128        313        640:  29%|██▊       | 231/809 [04:32<11:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101     0.4523     0.1128        241        640:  29%|██▊       | 232/809 [04:33<11:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4533     0.1126         60        640:  29%|██▉       | 234/809 [04:35<11:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4533     0.1126        283        640:  29%|██▉       | 236/809 [04:38<11:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4532     0.1126        483        640:  29%|██▉       | 237/809 [04:39<11:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.096      0.454     0.1124        157        640:  30%|██▉       | 239/809 [04:41<11:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.097     0.4537     0.1123        251        640:  30%|██▉       | 241/809 [04:44<11:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.096     0.4542     0.1122        229        640:  30%|███       | 243/809 [04:46<11:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.096     0.4541     0.1122        227        640:  30%|███       | 244/809 [04:47<11:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.094     0.4549     0.1121        172        640:  30%|███       | 246/809 [04:50<10:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.094     0.4546     0.1119        187        640:  31%|███       | 248/809 [04:52<10:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.095     0.4543      0.112        325        640:  31%|███       | 249/809 [04:53<10:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4534     0.1127        271        640:  31%|███       | 251/809 [04:55<10:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4533     0.1126        143        640:  31%|███▏      | 253/809 [04:58<10:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4529     0.1127        171        640:  32%|███▏      | 255/809 [05:00<10:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099      0.453     0.1127        206        640:  32%|███▏      | 256/809 [05:01<11:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4534     0.1124        254        640:  32%|███▏      | 258/809 [05:04<10:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4532      0.113        188        640:  32%|███▏      | 260/809 [05:06<10:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4532     0.1131        346        640:  32%|███▏      | 261/809 [05:07<10:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.103     0.4527     0.1141        466        640:  33%|███▎      | 263/809 [05:10<10:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101     0.4531      0.114        267        640:  33%|███▎      | 265/809 [05:12<10:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4535     0.1138        184        640:  33%|███▎      | 266/809 [05:13<10:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4536     0.1137        254        640:  33%|███▎      | 268/809 [05:15<10:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4541     0.1137        232        640:  33%|███▎      | 270/809 [05:18<10:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4541     0.1138        383        640:  33%|███▎      | 271/809 [05:19<10:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4544     0.1137        103        640:  34%|███▎      | 273/809 [05:22<10:53,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.097     0.4548     0.1135        172        640:  34%|███▍      | 275/809 [05:24<10:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.097     0.4548     0.1135        381        640:  34%|███▍      | 276/809 [05:25<10:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.094     0.4553     0.1132        132        640:  34%|███▍      | 278/809 [05:27<10:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.094     0.4554     0.1131        214        640:  35%|███▍      | 280/809 [05:30<10:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.092     0.4561     0.1129        112        640:  35%|███▍      | 282/809 [05:32<10:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.092     0.4561     0.1127        243        640:  35%|███▍      | 283/809 [05:33<10:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.09     0.4569     0.1125        256        640:  35%|███▌      | 285/809 [05:36<10:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.089     0.4573     0.1123        223        640:  35%|███▌      | 287/809 [05:38<10:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.09     0.4568     0.1125        566        640:  36%|███▌      | 288/809 [05:39<10:40,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.093      0.456     0.1127        225        640:  36%|███▌      | 290/809 [05:42<10:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.094     0.4556     0.1128        177        640:  36%|███▌      | 292/809 [05:44<10:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.095     0.4556     0.1127        160        640:  36%|███▋      | 294/809 [05:46<09:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.097     0.4551     0.1128        241        640:  36%|███▋      | 295/809 [05:47<09:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.097     0.4551     0.1127        223        640:  37%|███▋      | 297/809 [05:50<10:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.097     0.4552     0.1128        319        640:  37%|███▋      | 299/809 [05:52<09:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4551      0.113        321        640:  37%|███▋      | 300/809 [05:53<09:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4551     0.1131        321        640:  37%|███▋      | 302/809 [05:56<09:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4548      0.113        441        640:  38%|███▊      | 304/809 [05:58<10:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4549     0.1129        100        640:  38%|███▊      | 305/809 [05:59<10:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4545     0.1129        308        640:  38%|███▊      | 307/809 [06:02<09:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101     0.4541     0.1131        493        640:  38%|███▊      | 309/809 [06:04<09:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101      0.454     0.1131         63        640:  38%|███▊      | 311/809 [06:06<09:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101     0.4538      0.113        364        640:  39%|███▊      | 312/809 [06:07<09:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099      0.454     0.1129        118        640:  39%|███▉      | 314/809 [06:10<09:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4539     0.1129        243        640:  39%|███▉      | 316/809 [06:12<09:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4541     0.1128        192        640:  39%|███▉      | 317/809 [06:13<09:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4539     0.1127        361        640:  39%|███▉      | 319/809 [06:16<09:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4541     0.1125        183        640:  40%|███▉      | 321/809 [06:18<09:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4541     0.1124        246        640:  40%|███▉      | 323/809 [06:20<09:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.099     0.4539     0.1124        240        640:  40%|████      | 324/809 [06:21<09:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4535     0.1122        215        640:  40%|████      | 326/809 [06:24<09:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4535      0.112        328        640:  41%|████      | 328/809 [06:26<09:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.098     0.4538      0.112        240        640:  41%|████      | 330/809 [06:28<09:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G        1.1     0.4535     0.1127        487        640:  41%|████      | 331/809 [06:29<09:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101     0.4533     0.1137        285        640:  41%|████      | 333/809 [06:32<09:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.101     0.4531     0.1136        120        640:  41%|████▏     | 335/809 [06:34<09:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.102      0.453     0.1137        505        640:  42%|████▏     | 336/809 [06:35<09:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.103     0.4523     0.1136        166        640:  42%|████▏     | 338/809 [06:38<09:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4519     0.1139        205        640:  42%|████▏     | 340/809 [06:40<09:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4519     0.1138        164        640:  42%|████▏     | 341/809 [06:41<09:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4513     0.1144        145        640:  42%|████▏     | 343/809 [06:44<09:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4511     0.1147        343        640:  43%|████▎     | 345/809 [06:46<09:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4509     0.1149        199        640:  43%|████▎     | 347/809 [06:48<08:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109      0.451     0.1148        151        640:  43%|████▎     | 348/809 [06:49<09:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109      0.451     0.1148        197        640:  43%|████▎     | 350/809 [06:52<08:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109      0.451     0.1146        213        640:  44%|████▎     | 352/809 [06:54<08:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4515     0.1145        113        640:  44%|████▍     | 354/809 [06:56<08:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4514     0.1145        203        640:  44%|████▍     | 355/809 [06:58<08:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4513     0.1145        233        640:  44%|████▍     | 357/809 [07:00<08:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4511     0.1147        235        640:  44%|████▍     | 359/809 [07:02<08:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4516     0.1145         92        640:  45%|████▍     | 361/809 [07:05<08:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4512     0.1151        253        640:  45%|████▍     | 363/809 [07:07<08:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4508     0.1154        438        640:  45%|████▍     | 364/809 [07:08<08:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4507     0.1153        171        640:  45%|████▌     | 366/809 [07:10<08:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4504     0.1154        391        640:  45%|████▌     | 368/809 [07:13<08:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112     0.4501     0.1161        452        640:  46%|████▌     | 369/809 [07:14<08:56,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4502     0.1159        219        640:  46%|████▌     | 371/809 [07:17<08:57,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4503     0.1158        256        640:  46%|████▌     | 372/809 [07:18<09:02,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4508     0.1156        255        640:  46%|████▌     | 374/809 [07:20<08:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4515     0.1155        139        640:  46%|████▋     | 376/809 [07:23<08:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4516     0.1155        246        640:  47%|████▋     | 378/809 [07:25<08:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4517     0.1154        163        640:  47%|████▋     | 379/809 [07:26<08:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4516     0.1153        352        640:  47%|████▋     | 381/809 [07:28<08:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.106     0.4519     0.1155        129        640:  47%|████▋     | 383/809 [07:31<08:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.106     0.4517     0.1155        307        640:  47%|████▋     | 384/809 [07:32<08:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105      0.452     0.1154        346        640:  48%|████▊     | 386/809 [07:34<08:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.106     0.4521     0.1156        483        640:  48%|████▊     | 388/809 [07:37<08:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4517     0.1158        361        640:  48%|████▊     | 389/809 [07:38<08:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4514     0.1157        225        640:  48%|████▊     | 391/809 [07:40<08:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4513     0.1163        189        640:  49%|████▊     | 393/809 [07:43<08:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4512     0.1162        299        640:  49%|████▉     | 395/809 [07:45<08:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108      0.451     0.1165        324        640:  49%|████▉     | 396/809 [07:46<08:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4508     0.1165        505        640:  49%|████▉     | 398/809 [07:49<08:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4513     0.1163        112        640:  49%|████▉     | 400/809 [07:51<08:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.106     0.4513     0.1163        235        640:  50%|████▉     | 401/809 [07:52<07:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4518     0.1162        307        640:  50%|████▉     | 403/809 [07:54<07:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4519      0.116        122        640:  50%|█████     | 405/809 [07:57<07:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.106     0.4514     0.1166        329        640:  50%|█████     | 407/809 [07:59<07:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4519     0.1165        118        640:  50%|█████     | 408/809 [08:00<07:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105      0.452     0.1165        154        640:  51%|█████     | 410/809 [08:03<07:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.104     0.4519     0.1165        289        640:  51%|█████     | 412/809 [08:05<07:37,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.104     0.4518     0.1164        266        640:  51%|█████     | 413/809 [08:06<07:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4515     0.1164        198        640:  51%|█████▏    | 415/809 [08:08<07:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.106     0.4512     0.1164        385        640:  52%|█████▏    | 417/809 [08:11<07:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4508     0.1172        218        640:  52%|█████▏    | 419/809 [08:13<07:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4508     0.1171        415        640:  52%|█████▏    | 420/809 [08:14<07:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4508      0.117        234        640:  52%|█████▏    | 422/809 [08:17<07:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.106     0.4513     0.1168        219        640:  52%|█████▏    | 424/809 [08:19<07:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4514     0.1168        235        640:  53%|█████▎    | 425/809 [08:20<07:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.105     0.4513     0.1167        238        640:  53%|█████▎    | 427/809 [08:22<07:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.104     0.4517     0.1166        233        640:  53%|█████▎    | 429/809 [08:25<07:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.106     0.4513     0.1167        380        640:  53%|█████▎    | 430/809 [08:26<07:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107      0.451     0.1168        153        640:  53%|█████▎    | 432/809 [08:28<07:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4506      0.117        240        640:  54%|█████▎    | 434/809 [08:31<07:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4507      0.117        135        640:  54%|█████▍    | 436/809 [08:33<07:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4507     0.1171        343        640:  54%|█████▍    | 437/809 [08:34<07:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4507      0.117        601        640:  54%|█████▍    | 439/809 [08:37<07:30,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107      0.451     0.1169        163        640:  55%|█████▍    | 441/809 [08:39<07:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4509     0.1168        190        640:  55%|█████▍    | 442/809 [08:40<07:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4503     0.1173        658        640:  55%|█████▍    | 444/809 [08:43<07:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.108     0.4507     0.1172        173        640:  55%|█████▌    | 446/809 [08:45<07:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4509     0.1172        197        640:  55%|█████▌    | 447/809 [08:46<07:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.107     0.4508     0.1171        297        640:  56%|█████▌    | 449/809 [08:49<07:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4505     0.1173        188        640:  56%|█████▌    | 451/809 [08:51<07:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4502     0.1176        580        640:  56%|█████▌    | 452/809 [08:52<07:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112     0.4494     0.1183        550        640:  56%|█████▌    | 454/809 [08:55<07:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.113     0.4493     0.1183        101        640:  56%|█████▋    | 456/809 [08:57<07:07,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.113     0.4492     0.1185        291        640:  56%|█████▋    | 457/809 [08:58<06:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114      0.449     0.1188        468        640:  57%|█████▋    | 459/809 [09:00<06:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4492     0.1186        213        640:  57%|█████▋    | 461/809 [09:03<06:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4491     0.1188        222        640:  57%|█████▋    | 462/809 [09:04<06:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4485     0.1191        298        640:  57%|█████▋    | 464/809 [09:06<06:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4485      0.119        312        640:  58%|█████▊    | 466/809 [09:09<06:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4483     0.1189        225        640:  58%|█████▊    | 468/809 [09:11<06:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.117      0.448     0.1191        619        640:  58%|█████▊    | 469/809 [09:13<06:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4482     0.1189        151        640:  58%|█████▊    | 471/809 [09:15<06:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4485     0.1187        246        640:  58%|█████▊    | 473/809 [09:17<06:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4486     0.1187        209        640:  59%|█████▊    | 474/809 [09:18<06:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4486     0.1186        151        640:  59%|█████▉    | 476/809 [09:21<06:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4481     0.1188        178        640:  59%|█████▉    | 478/809 [09:23<06:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4481     0.1187        132        640:  59%|█████▉    | 479/809 [09:24<06:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4484     0.1187        326        640:  59%|█████▉    | 481/809 [09:27<06:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4483     0.1188        332        640:  60%|█████▉    | 483/809 [09:29<06:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4483     0.1187        367        640:  60%|█████▉    | 484/809 [09:30<06:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4486     0.1186        145        640:  60%|██████    | 486/809 [09:33<06:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4487     0.1185        210        640:  60%|██████    | 488/809 [09:35<06:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4487     0.1184        194        640:  60%|██████    | 489/809 [09:36<06:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4489     0.1183        215        640:  61%|██████    | 491/809 [09:39<06:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4483     0.1187        361        640:  61%|██████    | 493/809 [09:41<06:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.117     0.4481     0.1189        118        640:  61%|██████    | 495/809 [09:43<06:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.118     0.4481      0.119        222        640:  61%|██████▏   | 496/809 [09:45<06:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.117     0.4483     0.1191        181        640:  62%|██████▏   | 498/809 [09:47<06:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.118      0.448     0.1192        518        640:  62%|██████▏   | 500/809 [09:49<05:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.118      0.448     0.1192        467        640:  62%|██████▏   | 502/809 [09:51<05:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.117     0.4481     0.1191        116        640:  62%|██████▏   | 504/809 [09:54<06:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.117      0.448      0.119        252        640:  62%|██████▏   | 505/809 [09:55<05:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4483      0.119        226        640:  63%|██████▎   | 507/809 [09:57<05:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4484      0.119        173        640:  63%|██████▎   | 509/809 [10:00<05:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.116     0.4486      0.119        374        640:  63%|██████▎   | 510/809 [10:01<05:45,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115      0.449     0.1189        145        640:  63%|██████▎   | 512/809 [10:03<05:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4491     0.1189        464        640:  64%|██████▎   | 514/809 [10:06<05:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4492     0.1188        280        640:  64%|██████▎   | 515/809 [10:07<05:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4492     0.1188        148        640:  64%|██████▍   | 517/809 [10:09<05:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4491     0.1188        368        640:  64%|██████▍   | 519/809 [10:12<05:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4491     0.1188        314        640:  64%|██████▍   | 520/809 [10:13<05:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4493     0.1186        178        640:  65%|██████▍   | 522/809 [10:15<05:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4493     0.1186        138        640:  65%|██████▍   | 524/809 [10:17<05:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4491     0.1186        364        640:  65%|██████▌   | 526/809 [10:20<05:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4491     0.1186        237        640:  65%|██████▌   | 527/809 [10:21<05:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114      0.449     0.1185        445        640:  65%|██████▌   | 529/809 [10:23<05:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4485     0.1186        388        640:  66%|██████▌   | 531/809 [10:26<05:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4487     0.1186        210        640:  66%|██████▌   | 533/809 [10:28<05:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4489     0.1185        156        640:  66%|██████▌   | 534/809 [10:29<05:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4487     0.1184        147        640:  66%|██████▋   | 536/809 [10:31<05:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4489     0.1183        368        640:  67%|██████▋   | 538/809 [10:34<05:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114      0.449     0.1183        215        640:  67%|██████▋   | 539/809 [10:35<05:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4488     0.1183        386        640:  67%|██████▋   | 541/809 [10:37<05:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4492     0.1182        126        640:  67%|██████▋   | 543/809 [10:40<05:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.113     0.4494     0.1181        261        640:  67%|██████▋   | 545/809 [10:42<05:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.113     0.4493     0.1181        285        640:  67%|██████▋   | 546/809 [10:43<05:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.113     0.4493     0.1181        245        640:  68%|██████▊   | 548/809 [10:46<05:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4491      0.118        123        640:  68%|██████▊   | 550/809 [10:48<05:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4488     0.1184        554        640:  68%|██████▊   | 551/809 [10:49<05:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.115     0.4487     0.1184        348        640:  68%|██████▊   | 553/809 [10:51<04:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114      0.449     0.1183        350        640:  69%|██████▊   | 555/809 [10:54<04:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4489     0.1182        193        640:  69%|██████▊   | 556/809 [10:55<05:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.114     0.4489     0.1183        206        640:  69%|██████▉   | 558/809 [10:57<05:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.113     0.4492     0.1182        157        640:  69%|██████▉   | 560/809 [11:00<05:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.113     0.4492     0.1182        299        640:  69%|██████▉   | 561/809 [11:01<04:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112     0.4495     0.1181        294        640:  70%|██████▉   | 563/809 [11:03<04:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112     0.4499      0.118        278        640:  70%|██████▉   | 565/809 [11:06<04:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4502      0.118        163        640:  70%|██████▉   | 566/809 [11:07<04:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4503     0.1183        131        640:  70%|███████   | 568/809 [11:09<04:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4504     0.1182        210        640:  70%|███████   | 570/809 [11:12<04:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4507     0.1181        362        640:  71%|███████   | 572/809 [11:14<04:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4509     0.1181        162        640:  71%|███████   | 573/809 [11:15<04:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4509     0.1181        487        640:  71%|███████   | 575/809 [11:18<04:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109      0.451      0.118        251        640:  71%|███████   | 576/809 [11:19<04:53,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4507     0.1186        264        640:  71%|███████▏  | 578/809 [11:21<04:43,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4508     0.1186        171        640:  72%|███████▏  | 580/809 [11:24<04:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4509     0.1186        154        640:  72%|███████▏  | 581/809 [11:25<04:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4506      0.119        456        640:  72%|███████▏  | 583/809 [11:27<04:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4502      0.119        312        640:  72%|███████▏  | 585/809 [11:30<04:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112       0.45     0.1193        484        640:  73%|███████▎  | 587/809 [11:32<04:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4503     0.1192        101        640:  73%|███████▎  | 588/809 [11:33<04:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4505     0.1191        302        640:  73%|███████▎  | 590/809 [11:36<04:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4509     0.1189        280        640:  73%|███████▎  | 592/809 [11:38<04:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4508     0.1189        212        640:  73%|███████▎  | 593/809 [11:39<04:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4508     0.1188        177        640:  74%|███████▎  | 595/809 [11:42<04:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4507     0.1188        426        640:  74%|███████▍  | 597/809 [11:44<04:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4508     0.1187        131        640:  74%|███████▍  | 599/809 [11:46<04:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4505     0.1187        403        640:  74%|███████▍  | 600/809 [11:47<04:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4506     0.1188        274        640:  74%|███████▍  | 602/809 [11:50<04:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4508     0.1188        117        640:  75%|███████▍  | 604/809 [11:52<03:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11      0.451     0.1187        126        640:  75%|███████▍  | 605/809 [11:53<03:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4514     0.1187        198        640:  75%|███████▌  | 607/809 [11:55<03:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4514     0.1188        190        640:  75%|███████▌  | 609/809 [11:58<03:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4516     0.1187        191        640:  76%|███████▌  | 611/809 [12:00<03:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4517     0.1187        216        640:  76%|███████▌  | 612/809 [12:01<03:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4518     0.1186        205        640:  76%|███████▌  | 614/809 [12:04<03:44,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4519     0.1186        387        640:  76%|███████▌  | 616/809 [12:06<03:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4519     0.1186        320        640:  76%|███████▋  | 618/809 [12:08<03:39,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11      0.452     0.1186        313        640:  77%|███████▋  | 619/809 [12:09<03:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4521     0.1186        302        640:  77%|███████▋  | 621/809 [12:12<03:37,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4519     0.1187        279        640:  77%|███████▋  | 623/809 [12:14<03:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4522     0.1185        154        640:  77%|███████▋  | 625/809 [12:16<03:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4523     0.1185        291        640:  77%|███████▋  | 626/809 [12:17<03:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4522     0.1186        483        640:  78%|███████▊  | 628/809 [12:20<03:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4521     0.1188        381        640:  78%|███████▊  | 630/809 [12:22<03:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4523     0.1187        292        640:  78%|███████▊  | 631/809 [12:23<03:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.109     0.4524     0.1187        264        640:  78%|███████▊  | 633/809 [12:26<03:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4522     0.1191        293        640:  78%|███████▊  | 635/809 [12:28<03:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4521      0.119        265        640:  79%|███████▊  | 636/809 [12:30<03:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G       1.11     0.4522      0.119        132        640:  79%|███████▉  | 638/809 [12:32<03:31,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4518     0.1191        478        640:  79%|███████▉  | 640/809 [12:34<03:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4519     0.1191        382        640:  79%|███████▉  | 641/809 [12:35<03:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112     0.4518     0.1191        318        640:  79%|███████▉  | 643/809 [12:38<03:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.113     0.4516     0.1198        384        640:  80%|███████▉  | 645/809 [12:40<03:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112     0.4519     0.1197        221        640:  80%|███████▉  | 647/809 [12:42<03:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112     0.4519     0.1196        304        640:  80%|████████  | 648/809 [12:44<03:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4521     0.1195        198        640:  80%|████████  | 650/809 [12:46<03:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4521     0.1195        366        640:  81%|████████  | 652/809 [12:48<03:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112      0.452     0.1196        320        640:  81%|████████  | 654/809 [12:50<02:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112      0.452     0.1196        200        640:  81%|████████  | 656/809 [12:53<02:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4521     0.1195        277        640:  81%|████████  | 657/809 [12:54<02:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.111     0.4523     0.1194        312        640:  81%|████████▏ | 659/809 [12:56<02:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      13.5G      1.112      0.452     0.1197        481        640:  82%|████████▏ | 660/809 [12:57<02:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.114     0.4516     0.1214        326        640:  82%|████████▏ | 662/809 [13:00<03:09,  1.29s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113      0.452     0.1213         97        640:  82%|████████▏ | 664/809 [13:03<02:58,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4521     0.1213        209        640:  82%|████████▏ | 665/809 [13:04<02:55,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4519     0.1214        372        640:  82%|████████▏ | 667/809 [13:06<02:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.114     0.4519     0.1214        217        640:  83%|████████▎ | 669/809 [13:09<02:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.114      0.452     0.1214        161        640:  83%|██���█████▎ | 671/809 [13:11<02:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4522     0.1213        210        640:  83%|████████▎ | 672/809 [13:12<02:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4524     0.1212        204        640:  83%|████████▎ | 674/809 [13:14<02:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4525     0.1212        201        640:  84%|████████▎ | 676/809 [13:17<02:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4525     0.1212        139        640:  84%|████████▎ | 677/809 [13:18<02:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4526      0.121         78        640:  84%|████████▍ | 679/809 [13:20<02:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4525      0.121        281        640:  84%|████████▍ | 681/809 [13:23<02:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4526     0.1209        257        640:  84%|████████▍ | 683/809 [13:25<02:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4526     0.1209        365        640:  85%|████████▍ | 684/809 [13:26<02:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4525      0.121        329        640:  85%|████████▍ | 686/809 [13:28<02:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4526      0.121        552        640:  85%|████████▌ | 688/809 [13:31<02:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4529     0.1209         98        640:  85%|████████▌ | 689/809 [13:32<02:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4529      0.121        387        640:  85%|████████▌ | 691/809 [13:34<02:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4529     0.1209        266        640:  86%|████████▌ | 693/809 [13:37<02:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4529     0.1209        340        640:  86%|████████▌ | 694/809 [13:38<02:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113      0.453     0.1212        182        640:  86%|████████▌ | 696/809 [13:40<02:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4532     0.1211        238        640:  86%|████████▋ | 698/809 [13:43<02:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4533      0.121        212        640:  87%|████████▋ | 700/809 [13:45<02:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4533      0.121        409        640:  87%|████████▋ | 701/809 [13:46<02:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4536     0.1209        159        640:  87%|████████▋ | 703/809 [13:48<02:02,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4532      0.121        444        640:  87%|████████▋ | 705/809 [13:51<02:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4531      0.121        336        640:  87%|████████▋ | 706/809 [13:52<02:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.114     0.4528     0.1209        289        640:  88%|████████▊ | 708/809 [13:54<01:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.115     0.4526      0.121        276        640:  88%|████████▊ | 710/809 [13:57<01:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4523     0.1212        342        640:  88%|████████▊ | 712/809 [13:59<01:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.115     0.4524     0.1211        264        640:  88%|████████▊ | 713/809 [14:00<01:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.115     0.4526      0.121        260        640:  88%|████████▊ | 715/809 [14:02<01:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4525     0.1213        162        640:  89%|████████▊ | 717/809 [14:05<01:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4524     0.1213        360        640:  89%|████████▉ | 718/809 [14:06<01:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4522     0.1214        397        640:  89%|████████▉ | 720/809 [14:08<01:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4523     0.1214        197        640:  89%|████████▉ | 722/809 [14:11<01:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4524     0.1213        182        640:  89%|████████▉ | 723/809 [14:12<01:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.118      0.452     0.1218        605        640:  90%|████████▉ | 725/809 [14:14<01:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.118      0.452     0.1217        158        640:  90%|████████▉ | 727/809 [14:17<01:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.118     0.4521     0.1217        204        640:  90%|████████▉ | 728/809 [14:18<01:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.118     0.4522     0.1215        248        640:  90%|█████████ | 730/809 [14:20<01:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4524     0.1215        196        640:  90%|█████████ | 732/809 [14:23<01:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4527     0.1213        100        640:  91%|█████████ | 734/809 [14:25<01:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.115     0.4528     0.1212         98        640:  91%|█████████ | 735/809 [14:26<01:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.115     0.4529     0.1211        300        640:  91%|█████████ | 737/809 [14:28<01:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4527     0.1214        384        640:  91%|█████████▏| 739/809 [14:31<01:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4528     0.1213        122        640:  91%|█████████▏| 740/809 [14:32<01:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4527     0.1213        314        640:  92%|█████████▏| 742/809 [14:34<01:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4526     0.1213        263        640:  92%|█████████▏| 744/809 [14:37<01:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4525     0.1215        302        640:  92%|█████████▏| 746/809 [14:39<01:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4524     0.1215        304        640:  92%|█████████▏| 747/809 [14:40<01:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4525     0.1215        222        640:  93%|█████████▎| 749/809 [14:43<01:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4527     0.1215        170        640:  93%|█████████▎| 751/809 [14:45<01:06,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4525     0.1217        416        640:  93%|█████████▎| 752/809 [14:46<01:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4526     0.1216        465        640:  93%|█████████▎| 754/809 [14:49<01:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4524     0.1217        347        640:  93%|█████████▎| 756/809 [14:51<01:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4523     0.1217        335        640:  94%|█████████▎| 758/809 [14:53<00:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4523     0.1216        359        640:  94%|█████████▍| 759/809 [14:54<00:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.117     0.4525     0.1215        274        640:  94%|█████████▍| 761/809 [14:57<00:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4526     0.1214        237        640:  94%|█████████▍| 763/809 [14:59<00:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4527     0.1214         60        640:  94%|█████████▍| 764/809 [15:00<00:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4526     0.1213        306        640:  95%|█████████▍| 766/809 [15:03<00:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4526     0.1213        168        640:  95%|█████████▍| 768/809 [15:05<00:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4526     0.1212        279        640:  95%|█████████▌| 770/809 [15:07<00:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4527     0.1212        113        640:  95%|█████████▌| 771/809 [15:08<00:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4527     0.1211        132        640:  96%|█████████▌| 773/809 [15:11<00:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4527     0.1211        383        640:  96%|█████████▌| 775/809 [15:13<00:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4528      0.121        266        640:  96%|█████████▌| 776/809 [15:14<00:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.115     0.4527     0.1209        347        640:  96%|█████████▌| 778/809 [15:17<00:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4526     0.1209        227        640:  96%|█████████▋| 780/809 [15:19<00:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4524     0.1211        451        640:  97%|█████████▋| 781/809 [15:20<00:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.116     0.4526      0.121        197        640:  97%|█████████▋| 783/809 [15:23<00:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.115     0.4528     0.1209        184        640:  97%|█████████▋| 785/809 [15:25<00:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.115      0.453     0.1208        257        640:  97%|█████████▋| 787/809 [15:27<00:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.114     0.4532     0.1208         86        640:  97%|█████████▋| 788/809 [15:28<00:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.114     0.4533     0.1207        128        640:  98%|█████████▊| 790/809 [15:31<00:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.114     0.4533     0.1206        282        640:  98%|█████████▊| 792/809 [15:33<00:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4535     0.1205        127        640:  98%|█████████▊| 794/809 [15:35<00:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4536     0.1204        190        640:  98%|█████████▊| 795/809 [15:37<00:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4537     0.1203        286        640:  99%|█████████▊| 797/809 [15:39<00:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4537     0.1203        333        640:  99%|█████████▉| 799/809 [15:41<00:11,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.112     0.4537     0.1203        314        640:  99%|█████████▉| 800/809 [15:42<00:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4535     0.1204        385        640:  99%|█████████▉| 802/809 [15:45<00:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4535     0.1203        263        640:  99%|█████████▉| 804/809 [15:47<00:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4535     0.1203        278        640: 100%|█████████▉| 806/809 [15:49<00:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     78/300      14.5G      1.113     0.4535     0.1202        199        640: 100%|█████████▉| 808/809 [15:52<00:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   6%|▌         | 4/69 [00:00<00:15,  4.22it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  17%|█▋        | 12/69 [00:02<00:13,  4.09it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  30%|███       | 21/69 [00:05<00:11,  4.09it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  42%|████▏     | 29/69 [00:07<00:09,  4.13it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  54%|█████▎    | 37/69 [00:09<00:07,  4.09it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  64%|██████▍   | 44/69 [00:10<00:06,  3.91it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  75%|███████▌  | 52/69 [00:12<00:04,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  88%|████████▊ | 61/69 [00:15<00:01,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                   all        548      38759      0.531      0.361      0.367       0.21
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward passinit__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.024     0.4705    0.08071        259        640:   0%|          | 2/809 [00:02<15:35,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.066     0.4661     0.0901        307        640:   0%|          | 4/809 [00:04<15:27,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.026     0.4838    0.08699        359        640:   1%|          | 6/809 [00:06<15:25,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.05      0.482     0.0898        244        640:   1%|          | 7/809 [00:08<15:42,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.032     0.4829     0.0905        143        640:   1%|          | 9/809 [00:10<15:40,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.062     0.4652    0.09097        136        640:   1%|▏         | 11/809 [00:12<15:58,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4643     0.1009        374        640:   1%|▏         | 12/809 [00:14<15:44,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.124     0.4527     0.1111        641        640:   2%|▏         | 14/809 [00:16<15:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4697      0.107        234        640:   2%|▏         | 16/809 [00:18<15:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.098     0.4641     0.1088        412        640:   2%|▏         | 17/809 [00:20<15:31,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.096      0.463     0.1073        336        640:   2%|▏         | 19/809 [00:22<15:35,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077       0.47     0.1039         95        640:   3%|▎         | 21/809 [00:24<15:47,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.066     0.4691      0.103        198        640:   3%|▎         | 22/809 [00:26<15:37,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.05     0.4751     0.1018        252        640:   3%|▎         | 24/809 [00:28<15:41,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074     0.4674     0.1059        441        640:   3%|▎         | 26/809 [00:30<15:31,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077     0.4662     0.1057        292        640:   3%|▎         | 27/809 [00:31<15:24,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.069     0.4678     0.1043        207        640:   4%|▎         | 29/809 [00:34<15:21,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077     0.4661     0.1043        270        640:   4%|▍         | 31/809 [00:36<15:46,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.069     0.4673     0.1032        265        640:   4%|▍         | 32/809 [00:37<15:26,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4673     0.1034        216        640:   4%|▍         | 34/809 [00:40<15:04,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.058     0.4736     0.1025        266        640:   4%|▍         | 36/809 [00:42<15:05,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.045     0.4757     0.1014        173        640:   5%|▍         | 38/809 [00:44<14:57,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.044     0.4759     0.1014        237        640:   5%|▍         | 39/809 [00:46<15:15,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.061     0.4731     0.1077        362        640:   5%|▌         | 41/809 [00:48<15:05,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073      0.469     0.1089        275        640:   5%|▌         | 43/809 [00:50<14:45,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.067     0.4731     0.1085        252        640:   5%|▌         | 44/809 [00:51<14:47,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.065     0.4738     0.1076        255        640:   6%|▌         | 46/809 [00:54<14:50,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4711     0.1073        319        640:   6%|▌         | 48/809 [00:56<14:50,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.076     0.4712     0.1103        625        640:   6%|▌         | 50/809 [00:59<14:48,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075      0.471     0.1102        277        640:   6%|▋         | 51/809 [01:00<14:59,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.067      0.474     0.1089        179        640:   7%|▋         | 53/809 [01:02<14:44,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074     0.4721     0.1108        495        640:   7%|▋         | 55/809 [01:05<15:14,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.078      0.471     0.1108        263        640:   7%|▋         | 56/809 [01:06<15:01,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.086     0.4673     0.1114        333        640:   7%|▋         | 58/809 [01:08<14:37,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.094     0.4655     0.1143        213        640:   7%|▋         | 60/809 [01:10<14:51,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.089      0.467      0.114        117        640:   8%|▊         | 61/809 [01:12<14:45,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.088     0.4676     0.1144        386        640:   8%|▊         | 63/809 [01:14<14:46,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.092     0.4669     0.1164        225        640:   8%|▊         | 65/809 [01:16<14:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.079     0.4708     0.1148        172        640:   8%|▊         | 67/809 [01:19<14:22,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.076     0.4712     0.1142        265        640:   8%|▊         | 68/809 [01:20<14:16,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.087     0.4676      0.117        464        640:   9%|▊         | 70/809 [01:22<14:34,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.081     0.4694     0.1165        211        640:   9%|▉         | 72/809 [01:25<14:31,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.081     0.4694     0.1162        299        640:   9%|▉         | 73/809 [01:26<14:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4714     0.1151        200        640:   9%|▉         | 75/809 [01:28<14:24,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4721     0.1155        184        640:  10%|▉         | 77/809 [01:30<14:03,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4719     0.1148        292        640:  10%|▉         | 79/809 [01:33<14:09,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4722     0.1148        151        640:  10%|▉         | 80/809 [01:34<14:04,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.067     0.4733     0.1137        135        640:  10%|█         | 82/809 [01:36<14:08,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.065     0.4736     0.1136        265        640:  10%|█         | 84/809 [01:38<13:55,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.064     0.4735     0.1133        299        640:  11%|█         | 86/809 [01:41<13:45,  1.14s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.063     0.4735     0.1129        162        640:  11%|█         | 87/809 [01:42<14:01,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.069     0.4711     0.1133        190        640:  11%|█         | 89/809 [01:44<13:50,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4695     0.1137        282        640:  11%|█         | 91/809 [01:47<13:47,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.076     0.4685     0.1138        323        640:  11%|█▏        | 92/809 [01:48<13:48,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.082     0.4661     0.1135        232        640:  12%|█▏        | 94/809 [01:50<13:43,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.085     0.4649      0.114        216        640:  12%|█▏        | 96/809 [01:52<13:46,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.08      0.466     0.1132        151        640:  12%|█▏        | 98/809 [01:55<13:36,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4653     0.1133        349        640:  12%|█▏        | 99/809 [01:56<13:50,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.081     0.4653     0.1129        118        640:  12%|█▏        | 101/809 [01:58<13:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077     0.4664     0.1118         88        640:  13%|█▎        | 103/809 [02:01<13:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.08     0.4656     0.1119        325        640:  13%|█▎        | 104/809 [02:02<13:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.081     0.4651     0.1117        148        640:  13%|█▎        | 106/809 [02:04<13:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.084     0.4642     0.1114        418        640:  13%|█▎        | 108/809 [02:06<13:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.088     0.4623     0.1114        237        640:  14%|█▎        | 110/809 [02:09<13:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.09     0.4616     0.1113        262        640:  14%|█▎        | 111/809 [02:10<13:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.091     0.4611     0.1108        193        640:  14%|█▍        | 113/809 [02:12<13:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.095     0.4599      0.111        242        640:  14%|█▍        | 115/809 [02:15<13:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.09     0.4607     0.1104        241        640:  14%|█▍        | 117/809 [02:17<13:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.091     0.4606     0.1102        242        640:  15%|█▍        | 118/809 [02:18<13:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.089      0.461     0.1103        253        640:  15%|█▍        | 120/809 [02:20<13:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.087     0.4615       0.11        102        640:  15%|█▌        | 122/809 [02:23<13:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.09     0.4604     0.1101        469        640:  15%|█▌        | 123/809 [02:24<13:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.093     0.4598     0.1103        286        640:  15%|█▌        | 125/809 [02:26<13:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.092     0.4604     0.1102        150        640:  16%|█▌        | 127/809 [02:29<13:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.094     0.4602     0.1104        250        640:  16%|█▌        | 129/809 [02:31<13:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.095     0.4602     0.1106        263        640:  16%|█▌        | 130/809 [02:32<13:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.092     0.4614     0.1102        163        640:  16%|█▋        | 132/809 [02:34<13:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.094     0.4616     0.1103        166        640:  17%|█▋        | 134/809 [02:37<13:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.092     0.4625     0.1101        131        640:  17%|█▋        | 135/809 [02:38<13:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.088      0.463     0.1097        192        640:  17%|█▋        | 137/809 [02:40<13:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.088     0.4628     0.1092         70        640:  17%|█▋        | 139/809 [02:43<13:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.086     0.4629     0.1088        225        640:  17%|█▋        | 141/809 [02:45<12:48,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.088     0.4621     0.1086        380        640:  18%|█▊        | 142/809 [02:46<13:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.087     0.4626     0.1083        355        640:  18%|█▊        | 144/809 [02:48<12:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.088     0.4629     0.1087        569        640:  18%|█▊        | 146/809 [02:51<12:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.087     0.4635     0.1087        106        640:  18%|█▊        | 147/809 [02:52<12:42,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.086     0.4636     0.1083        208        640:  18%|█▊        | 149/809 [02:54<12:35,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.086     0.4639     0.1084        256        640:  19%|█▊        | 151/809 [02:56<12:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4648      0.108        207        640:  19%|█▉        | 153/809 [02:59<12:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.085     0.4649     0.1089        152        640:  19%|█▉        | 155/809 [03:01<12:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.082     0.4656     0.1087         63        640:  19%|█▉        | 157/809 [03:03<12:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.08     0.4663     0.1084        142        640:  20%|█▉        | 158/809 [03:05<12:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.078     0.4672     0.1082        199        640:  20%|█▉        | 160/809 [03:07<12:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077     0.4672     0.1079        367        640:  20%|██        | 162/809 [03:09<12:27,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077     0.4675     0.1077        248        640:  20%|██        | 163/809 [03:10<12:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074     0.4679     0.1073        190        640:  20%|██        | 165/809 [03:13<12:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075      0.468     0.1074        133        640:  21%|██        | 167/809 [03:15<12:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4684     0.1072         52        640:  21%|██        | 169/809 [03:18<12:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4689     0.1071        208        640:  21%|██        | 170/809 [03:19<12:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.068     0.4692      0.107        260        640:  21%|██▏       | 172/809 [03:21<12:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.065     0.4699     0.1066        184        640:  22%|██▏       | 174/809 [03:23<12:09,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.064     0.4702     0.1067        330        640:  22%|██▏       | 175/809 [03:24<12:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.067     0.4693     0.1072        619        640:  22%|██▏       | 177/809 [03:27<12:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.067      0.469     0.1074        423        640:  22%|██▏       | 179/809 [03:29<12:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.068      0.469     0.1073        274        640:  22%|██▏       | 181/809 [03:31<12:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.066     0.4693     0.1072        205        640:  22%|██▏       | 182/809 [03:33<12:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.067     0.4691     0.1085         92        640:  23%|██▎       | 184/809 [03:35<12:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4681     0.1086        280        640:  23%|██▎       | 186/809 [03:37<12:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4679     0.1087        418        640:  23%|██▎       | 187/809 [03:39<12:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4666     0.1097        677        640:  23%|██▎       | 189/809 [03:41<12:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4668     0.1099        294        640:  24%|██▎       | 191/809 [03:43<12:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4667       0.11        414        640:  24%|██▎       | 192/809 [03:45<12:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4672     0.1098        292        640:  24%|██▍       | 194/809 [03:47<12:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4665       0.11        388        640:  24%|██▍       | 196/809 [03:49<11:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4669     0.1096        215        640:  24%|██▍       | 198/809 [03:52<11:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4674     0.1094        184        640:  25%|██▍       | 199/809 [03:53<12:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.069      0.468     0.1091        140        640:  25%|██▍       | 201/809 [03:55<11:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.067     0.4683      0.109        140        640:  25%|██▌       | 203/809 [03:57<11:37,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.067     0.4683     0.1089        275        640:  25%|██▌       | 204/809 [03:59<11:37,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.064     0.4686     0.1086        304        640:  25%|██▌       | 206/809 [04:01<11:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.063      0.469     0.1085        183        640:  26%|██▌       | 208/809 [04:03<11:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.064     0.4687     0.1096        178        640:  26%|██▌       | 210/809 [04:06<11:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.065     0.4685     0.1097        422        640:  26%|██▌       | 211/809 [04:07<11:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.065     0.4685     0.1096        307        640:  26%|���█▋       | 213/809 [04:09<11:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.063     0.4688     0.1093        161        640:  27%|██▋       | 215/809 [04:11<11:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.062     0.4694     0.1092        209        640:  27%|██▋       | 216/809 [04:13<11:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.061     0.4696     0.1092        167        640:  27%|██▋       | 218/809 [04:15<11:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.062     0.4693     0.1092        229        640:  27%|██▋       | 220/809 [04:17<11:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.064     0.4689     0.1101        430        640:  27%|██▋       | 221/809 [04:19<11:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.064     0.4691     0.1101        273        640:  28%|██▊       | 223/809 [04:21<11:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.064     0.4691     0.1102        309        640:  28%|██▊       | 225/809 [04:23<11:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.063     0.4695     0.1102        160        640:  28%|██▊       | 227/809 [04:26<11:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.063     0.4696     0.1102        137        640:  28%|██▊       | 228/809 [04:27<11:09,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.062     0.4697     0.1103        210        640:  28%|██▊       | 230/809 [04:29<11:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.062     0.4697     0.1101        153        640:  29%|██▊       | 232/809 [04:31<11:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.063     0.4694     0.1102        172        640:  29%|██▉       | 234/809 [04:34<11:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.065     0.4689     0.1103        204        640:  29%|██▉       | 235/809 [04:35<11:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.067      0.468     0.1104        360        640:  29%|██▉       | 237/809 [04:37<11:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4668      0.111        243        640:  30%|██▉       | 239/809 [04:40<11:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07      0.467     0.1109        222        640:  30%|██▉       | 240/809 [04:41<10:55,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.069     0.4677     0.1107        245        640:  30%|██▉       | 242/809 [04:43<10:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4677      0.111        176        640:  30%|███       | 244/809 [04:45<10:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4671     0.1116        315        640:  30%|███       | 245/809 [04:47<11:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4669     0.1119        221        640:  31%|███       | 247/809 [04:49<11:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074     0.4666     0.1119        223        640:  31%|███       | 249/809 [04:51<11:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4671     0.1118        181        640:  31%|███       | 251/809 [04:54<10:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4675     0.1118        180        640:  31%|███       | 252/809 [04:55<10:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4678     0.1118        122        640:  31%|███▏      | 254/809 [04:57<10:40,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4681     0.1117        234        640:  32%|███▏      | 256/809 [05:00<10:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4682     0.1116         75        640:  32%|███▏      | 258/809 [05:02<10:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4681     0.1115        253        640:  32%|███▏      | 259/809 [05:03<10:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4678     0.1116        412        640:  32%|███▏      | 261/809 [05:05<10:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4677     0.1114        240        640:  33%|███▎      | 263/809 [05:08<10:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4676     0.1113        186        640:  33%|███▎      | 264/809 [05:09<10:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4675     0.1114        301        640:  33%|███▎      | 266/809 [05:11<10:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4677     0.1114        238        640:  33%|███▎      | 268/809 [05:13<10:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4677     0.1114        343        640:  33%|███▎      | 270/809 [05:16<10:22,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4677     0.1114        196        640:  33%|███▎      | 271/809 [05:17<10:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4674     0.1113        223        640:  34%|███▎      | 273/809 [05:19<10:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4671     0.1112        228        640:  34%|███▍      | 275/809 [05:22<10:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4675     0.1111         74        640:  34%|███▍      | 276/809 [05:23<10:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4672     0.1112        308        640:  34%|███▍      | 278/809 [05:25<10:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4674      0.111        230        640:  35%|███▍      | 280/809 [05:28<10:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4675     0.1109        154        640:  35%|███▍      | 281/809 [05:29<10:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4673     0.1109        212        640:  35%|███▍      | 283/809 [05:31<10:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074     0.4666     0.1116        384        640:  35%|███▌      | 285/809 [05:34<10:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4665     0.1115        358        640:  35%|███▌      | 286/809 [05:35<10:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4664     0.1114        168        640:  36%|███▌      | 288/809 [05:37<10:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4667     0.1111         59        640:  36%|███▌      | 290/809 [05:39<10:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4663     0.1122        571        640:  36%|███▌      | 292/809 [05:42<10:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4661     0.1123        330        640:  36%|███▌      | 293/809 [05:43<10:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077     0.4655     0.1128        234        640:  36%|███▋      | 295/809 [05:45<10:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.078     0.4653     0.1132        173        640:  37%|███▋      | 297/809 [05:48<10:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.079     0.4653     0.1135        317        640:  37%|███▋      | 298/809 [05:49<10:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.08      0.465     0.1137        484        640:  37%|███▋      | 300/809 [05:51<10:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.081     0.4648      0.114        449        640:  37%|███▋      | 302/809 [05:54<09:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.081     0.4646     0.1139        137        640:  37%|███▋      | 303/809 [05:55<10:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.081     0.4645     0.1139        321        640:  38%|███▊      | 305/809 [05:57<09:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.079      0.465     0.1136        191        640:  38%|███▊      | 307/809 [06:00<09:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.079     0.4649     0.1142        148        640:  38%|███▊      | 309/809 [06:02<09:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.081     0.4645     0.1148        579        640:  38%|███▊      | 310/809 [06:03<09:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.079     0.4651     0.1147        144        640:  39%|███▊      | 312/809 [06:06<09:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.079     0.4648     0.1148        314        640:  39%|███▉      | 314/809 [06:08<09:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.079      0.465     0.1149        212        640:  39%|███▉      | 316/809 [06:10<09:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077     0.4652     0.1147        165        640:  39%|███▉      | 318/809 [06:12<09:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.079     0.4648      0.115        446        640:  39%|███▉      | 319/809 [06:14<09:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.08     0.4644     0.1152        178        640:  40%|███▉      | 321/809 [06:16<09:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.082     0.4641     0.1153        334        640:  40%|███▉      | 323/809 [06:18<09:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4637     0.1154        453        640:  40%|████      | 324/809 [06:19<09:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4639     0.1153        398        640:  40%|████      | 326/809 [06:22<09:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4639     0.1151        183        640:  41%|████      | 328/809 [06:24<09:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4638     0.1152        273        640:  41%|████      | 330/809 [06:27<09:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4636     0.1154        225        640:  41%|████      | 331/809 [06:28<09:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4634     0.1153        393        640:  41%|████      | 333/809 [06:30<09:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4634     0.1151        309        640:  41%|████▏     | 335/809 [06:32<09:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.085     0.4628     0.1155        437        640:  42%|████▏     | 336/809 [06:34<09:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.085     0.4625     0.1154        124        640:  42%|████▏     | 338/809 [06:36<09:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.086     0.4624     0.1156        498        640:  42%|████▏     | 340/809 [06:38<09:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.086     0.4622     0.1155        318        640:  42%|████▏     | 342/809 [06:41<09:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.086      0.462     0.1157        393        640:  42%|████▏     | 343/809 [06:42<09:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.086     0.4619     0.1158        108        640:  43%|████▎     | 345/809 [06:44<09:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.086     0.4618     0.1156        265        640:  43%|████▎     | 347/809 [06:47<09:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.084      0.462     0.1155         97        640:  43%|████▎     | 348/809 [06:48<09:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.084     0.4621     0.1153        394        640:  43%|████▎     | 350/809 [06:50<08:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.084      0.462     0.1155        347        640:  44%|████▎     | 352/809 [06:52<08:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4621     0.1153        135        640:  44%|████▍     | 354/809 [06:55<08:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.081     0.4624     0.1153         34        640:  44%|████▍     | 355/809 [06:56<08:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.082     0.4623     0.1152        257        640:  44%|████▍     | 357/809 [06:58<08:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.083     0.4621     0.1153        295        640:  44%|████▍     | 359/809 [07:01<08:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.082     0.4622     0.1151        148        640:  44%|████▍     | 360/809 [07:02<08:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.081     0.4627      0.115         92        640:  45%|████▍     | 362/809 [07:04<08:34,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.08     0.4632     0.1149        256        640:  45%|████▍     | 364/809 [07:06<08:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.079     0.4637     0.1149        127        640:  45%|████▌     | 366/809 [07:09<08:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.078     0.4639     0.1148        102        640:  45%|████▌     | 367/809 [07:10<08:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077      0.464     0.1146        305        640:  46%|████▌     | 369/809 [07:12<08:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4644     0.1144         99        640:  46%|████▌     | 371/809 [07:15<08:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.076     0.4643     0.1145        361        640:  46%|████▌     | 372/809 [07:16<08:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.076     0.4643     0.1147        184        640:  46%|████▌     | 374/809 [07:18<08:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.076     0.4644     0.1145        202        640:  46%|████▋     | 376/809 [07:20<08:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4647     0.1144        114        640:  47%|████▋     | 378/809 [07:23<08:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.076     0.4645     0.1145        260        640:  47%|████▋     | 379/809 [07:24<08:12,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.076     0.4643     0.1144        351        640:  47%|████▋     | 381/809 [07:26<08:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077      0.464     0.1144        255        640:  47%|████▋     | 383/809 [07:28<08:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4642     0.1144         50        640:  47%|████▋     | 384/809 [07:30<08:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4642     0.1142        298        640:  48%|████▊     | 386/809 [07:32<08:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074     0.4645     0.1141         89        640:  48%|████▊     | 388/809 [07:34<08:38,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074     0.4646      0.114        219        640:  48%|████▊     | 389/809 [07:36<08:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4647     0.1139        497        640:  48%|████▊     | 391/809 [07:38<08:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4647     0.1138        358        640:  49%|████▊     | 393/809 [07:40<08:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4647     0.1141        433        640:  49%|████▉     | 395/809 [07:43<08:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4647     0.1141        273        640:  49%|████▉     | 396/809 [07:44<08:35,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072      0.465     0.1139        197        640:  49%|████▉     | 398/809 [07:46<08:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4653     0.1139        143        640:  49%|████▉     | 399/809 [07:48<08:23,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4657     0.1138        163        640:  50%|████▉     | 401/809 [07:50<08:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4655     0.1138        244        640:  50%|████▉     | 403/809 [07:52<07:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4656     0.1136        215        640:  50%|█████     | 405/809 [07:55<07:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4657     0.1136        299        640:  50%|███��█     | 406/809 [07:56<07:44,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4653     0.1136        440        640:  50%|█████     | 408/809 [07:58<07:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4652     0.1137        190        640:  51%|█████     | 410/809 [08:01<07:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4653     0.1136        247        640:  51%|█████     | 411/809 [08:02<07:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4651     0.1139        325        640:  51%|█████     | 413/809 [08:04<07:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4652     0.1141        235        640:  51%|█████▏    | 415/809 [08:06<07:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074     0.4649     0.1141        470        640:  52%|█████▏    | 417/809 [08:09<07:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074     0.4649      0.114        185        640:  52%|█████▏    | 418/809 [08:10<07:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074     0.4649     0.1141        327        640:  52%|█████▏    | 420/809 [08:12<07:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074      0.465      0.114        348        640:  52%|█████▏    | 422/809 [08:15<07:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074      0.465     0.1139        265        640:  52%|█████▏    | 423/809 [08:16<07:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4648     0.1141        127        640:  53%|█████▎    | 425/809 [08:18<07:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4646     0.1141        136        640:  53%|█████▎    | 427/809 [08:20<07:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.076     0.4647      0.114        230        640:  53%|█████▎    | 429/809 [08:23<07:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.076     0.4645     0.1139        397        640:  53%|█████▎    | 430/809 [08:24<07:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077     0.4642      0.114        298        640:  53%|█████▎    | 432/809 [08:26<07:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.078      0.464      0.114        212        640:  54%|█████▎    | 434/809 [08:29<07:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.077     0.4641     0.1139         92        640:  54%|█████▍    | 435/809 [08:30<07:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4645     0.1137        261        640:  54%|█████▍    | 437/809 [08:32<07:27,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075     0.4647     0.1136        319        640:  54%|█████▍    | 439/809 [08:35<07:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.076     0.4646     0.1136        365        640:  54%|█████▍    | 440/809 [08:36<07:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.075      0.465     0.1134        358        640:  55%|█████▍    | 442/809 [08:38<07:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.074     0.4653     0.1132        245        640:  55%|█████▍    | 444/809 [08:41<07:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4653     0.1131        201        640:  55%|█████▌    | 446/809 [08:43<07:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.073     0.4656      0.113        152        640:  55%|█████▌    | 447/809 [08:44<07:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4657     0.1129        135        640:  56%|█████▌    | 449/809 [08:46<07:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4657     0.1127        198        640:  56%|█████▌    | 451/809 [08:49<07:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.072     0.4656     0.1127        291        640:  56%|█████▌    | 452/809 [08:50<06:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07      0.466     0.1126        303        640:  56%|█████▌    | 454/809 [08:52<06:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4662     0.1125        106        640:  56%|█████▋    | 456/809 [08:55<06:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4658     0.1127        213        640:  57%|█████▋    | 458/809 [08:57<06:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4659     0.1126        296        640:  57%|█████▋    | 459/809 [08:58<06:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4658     0.1129        207        640:  57%|█████▋    | 461/809 [09:00<06:41,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4658     0.1128        203        640:  57%|█████▋    | 463/809 [09:03<06:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4661     0.1127        192        640:  57%|█████▋    | 464/809 [09:04<06:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.069     0.4662     0.1126        135        640:  58%|█████▊    | 466/809 [09:06<06:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.069     0.4665     0.1125        204        640:  58%|█████▊    | 468/809 [09:09<06:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.068     0.4666     0.1124        220        640:  58%|█████▊    | 470/809 [09:11<06:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.067     0.4668     0.1123        245        640:  58%|█████▊    | 471/809 [09:12<06:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.067      0.467     0.1124        136        640:  58%|█████▊    | 473/809 [09:14<06:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.068     0.4667     0.1127        326        640:  59%|█████▊    | 475/809 [09:17<06:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4664     0.1129        528        640:  59%|█████▉    | 476/809 [09:18<06:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.069     0.4664     0.1127        142        640:  59%|█████▉    | 478/809 [09:20<06:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4663     0.1127        472        640:  59%|█████▉    | 480/809 [09:23<06:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4662     0.1128        254        640:  60%|█████▉    | 482/809 [09:25<06:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071      0.466     0.1128        310        640:  60%|█████▉    | 484/809 [09:27<06:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4661     0.1127        190        640:  60%|█████▉    | 485/809 [09:28<06:12,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.069     0.4665     0.1126        182        640:  60%|██████    | 487/809 [09:31<06:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.069     0.4665     0.1125        332        640:  60%|██████    | 489/809 [09:33<06:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4664     0.1126        413        640:  61%|██████    | 491/809 [09:35<06:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4664     0.1126        240        640:  61%|██████    | 492/809 [09:37<06:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4663     0.1126        224        640:  61%|██████    | 494/809 [09:39<06:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4662     0.1126        186        640:  61%|██████▏   | 496/809 [09:41<06:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4663     0.1126        261        640:  61%|██████▏   | 497/809 [09:42<06:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.069     0.4666     0.1125        212        640:  62%|██████▏   | 499/809 [09:45<05:57,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G       1.07     0.4665     0.1126        253        640:  62%|██████▏   | 501/809 [09:47<05:53,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4663     0.1125        145        640:  62%|██████▏   | 503/809 [09:49<05:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4662      0.113        293        640:  62%|██████▏   | 504/809 [09:51<05:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4662     0.1129        350        640:  63%|██████▎   | 506/809 [09:53<05:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      13.6G      1.071     0.4661     0.1129        341        640:  63%|██████▎   | 507/809 [09:54<05:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.073     0.4655      0.113        323        640:  63%|██████▎   | 509/809 [09:57<06:16,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.073     0.4656     0.1129        196        640:  63%|██████▎   | 511/809 [09:59<06:07,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.073     0.4654      0.113        162        640:  63%|██████▎   | 513/809 [10:02<05:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.074     0.4652     0.1131        289        640:  64%|██████▎   | 514/809 [10:03<05:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.075     0.4648     0.1132        787        640:  64%|██████▍   | 516/809 [10:05<06:07,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.076     0.4645     0.1133        271        640:  64%|██████▍   | 518/809 [10:08<05:54,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.076     0.4647     0.1132        329        640:  64%|██████▍   | 519/809 [10:09<05:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4641     0.1143        354        640:  64%|██████▍   | 521/809 [10:11<05:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4642     0.1141        120        640:  65%|██████▍   | 523/809 [10:14<05:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4642     0.1142        434        640:  65%|██████▍   | 524/809 [10:15<05:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4643     0.1142        354        640:  65%|██████▌   | 526/809 [10:17<05:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4643     0.1141        149        640:  65%|██████▌   | 528/809 [10:19<05:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4647      0.114        130        640:  66%|██████▌   | 530/809 [10:22<05:20,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4646      0.114        185        640:  66%|██████▌   | 531/809 [10:23<05:20,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4647     0.1139        198        640:  66%|██████▌   | 533/809 [10:25<05:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4646     0.1138        223        640:  66%|██████▌   | 535/809 [10:28<05:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4645     0.1138        239        640:  66%|██████▋   | 536/809 [10:29<05:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4643     0.1138        262        640:  67%|██████▋   | 538/809 [10:31<05:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4645     0.1137        145        640:  67%|██████▋   | 540/809 [10:33<05:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079      0.464      0.114        246        640:  67%|██████▋   | 542/809 [10:36<05:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4642     0.1139        195        640:  67%|██████▋   | 543/809 [10:37<05:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4643     0.1138        291        640:  67%|██████▋   | 545/809 [10:39<05:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4645     0.1137        194        640:  68%|██████▊   | 547/809 [10:42<05:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4645     0.1137        206        640:  68%|██████▊   | 548/809 [10:43<05:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4645     0.1138         97        640:  68%|██████▊   | 550/809 [10:45<05:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4647     0.1138         87        640:  68%|██████▊   | 552/809 [10:47<05:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4646     0.1137        390        640:  68%|██████▊   | 554/809 [10:50<04:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4645     0.1137        209        640:  69%|██████▊   | 555/809 [10:51<04:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079      0.464      0.114        744        640:  69%|██████▉   | 557/809 [10:53<05:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4639     0.1139        103        640:  69%|██████▉   | 559/809 [10:56<05:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4637      0.114        310        640:  69%|██████▉   | 560/809 [10:57<04:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4639     0.1139        276        640:  69%|██████▉   | 562/809 [10:59<04:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4636     0.1142        468        640:  70%|██████▉   | 564/809 [11:02<04:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4638     0.1141        210        640:  70%|██████▉   | 565/809 [11:03<04:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4637     0.1142        271        640:  70%|███████   | 567/809 [11:05<04:56,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4638     0.1141        331        640:  70%|███████   | 569/809 [11:08<04:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078      0.464      0.114        218        640:  70%|███████   | 570/809 [11:09<04:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4641      0.114        219        640:  71%|███████   | 572/809 [11:11<04:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4642     0.1139        403        640:  71%|███████   | 574/809 [11:14<04:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4639     0.1143        450        640:  71%|███████   | 575/809 [11:15<04:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4637     0.1143        568        640:  71%|███████▏  | 577/809 [11:17<04:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4637     0.1143        376        640:  72%|███████▏  | 579/809 [11:20<04:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.08     0.4635     0.1143        342        640:  72%|███████▏  | 581/809 [11:22<04:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.08     0.4633     0.1143        424        640:  72%|███████▏  | 582/809 [11:23<04:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.081     0.4629     0.1144        440        640:  72%|███████▏  | 584/809 [11:26<04:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.082     0.4628     0.1145        436        640:  72%|███████▏  | 586/809 [11:28<04:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.081     0.4628     0.1144        280        640:  73%|███████▎  | 587/809 [11:29<04:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.081     0.4628     0.1145        290        640:  73%|███████▎  | 589/809 [11:31<04:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.081     0.4628     0.1144        219        640:  73%|███████▎  | 591/809 [11:34<04:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.081     0.4629     0.1143        259        640:  73%|███████▎  | 592/809 [11:35<04:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4631     0.1143        170        640:  73%|███████▎  | 594/809 [11:37<04:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4633     0.1142        151        640:  74%|███████▎  | 596/809 [11:40<04:21,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4633     0.1142        331        640:  74%|███████▍  | 597/809 [11:41<04:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4632     0.1143        424        640:  74%|███████▍  | 599/809 [11:44<04:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4632     0.1142        312        640:  74%|███████▍  | 601/809 [11:46<04:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4632     0.1142        354        640:  74%|███████▍  | 602/809 [11:47<04:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.078     0.4632     0.1141        308        640:  75%|███████▍  | 604/809 [11:49<04:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4634      0.114        232        640:  75%|███████▍  | 606/809 [11:52<03:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.076     0.4637     0.1139        125        640:  75%|███████▌  | 607/809 [11:53<03:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.075     0.4639     0.1138        230        640:  75%|███████▌  | 609/809 [11:55<03:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.074      0.464     0.1137        286        640:  76%|███████▌  | 611/809 [11:58<03:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.074      0.464     0.1137        400        640:  76%|███████▌  | 613/809 [12:00<03:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.074      0.464     0.1137        236        640:  76%|███████▌  | 614/809 [12:01<03:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.074     0.4641     0.1137        113        640:  76%|███████▌  | 616/809 [12:03<03:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.073      0.464     0.1136         94        640:  76%|███████▋  | 618/809 [12:06<03:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.073     0.4641     0.1136        193        640:  77%|███████▋  | 619/809 [12:07<03:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.072     0.4643     0.1136        391        640:  77%|███████▋  | 621/809 [12:09<03:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.072     0.4644     0.1135        214        640:  77%|███████▋  | 623/809 [12:12<03:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.072     0.4643     0.1135        195        640:  77%|███████▋  | 624/809 [12:13<03:46,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.072     0.4644     0.1134        191        640:  77%|███████▋  | 626/809 [12:15<03:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.071     0.4645     0.1134        301        640:  78%|███████▊  | 628/809 [12:18<03:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.071     0.4646     0.1133        211        640:  78%|███████▊  | 630/809 [12:20<03:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.071     0.4644     0.1133        364        640:  78%|███████▊  | 631/809 [12:21<03:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.072     0.4644     0.1133        192        640:  78%|███████▊  | 633/809 [12:24<03:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.072     0.4643     0.1132        171        640:  78%|███████▊  | 635/809 [12:26<03:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.072     0.4642     0.1134        299        640:  79%|███████▊  | 636/809 [12:27<03:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.072     0.4644     0.1133        127        640:  79%|███████▉  | 638/809 [12:29<03:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.073     0.4642     0.1133        170        640:  79%|███████▉  | 640/809 [12:32<03:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.073     0.4643     0.1134        212        640:  79%|███████▉  | 642/809 [12:34<03:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.073     0.4643     0.1134        351        640:  80%|███████▉  | 644/809 [12:36<03:10,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.073     0.4641     0.1135        201        640:  80%|███████▉  | 645/809 [12:37<03:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.074     0.4641     0.1136        239        640:  80%|███████▉  | 647/809 [12:40<03:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.074     0.4641     0.1135        298        640:  80%|████████  | 649/809 [12:42<03:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.074     0.4641     0.1136        117        640:  80%|████████  | 651/809 [12:45<03:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.074     0.4642     0.1136        243        640:  81%|████████  | 652/809 [12:46<03:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.074     0.4642     0.1136        381        640:  81%|████████  | 654/809 [12:48<03:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.075     0.4636     0.1136        435        640:  81%|████████  | 656/809 [12:50<03:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.077     0.4632      0.114        437        640:  81%|████████  | 657/809 [12:52<03:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.079     0.4626     0.1142        106        640:  81%|████████▏ | 659/809 [12:54<02:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.081      0.462     0.1142        179        640:  82%|████████▏ | 661/809 [12:56<02:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.082     0.4617     0.1143        304        640:  82%|████████▏ | 662/809 [12:57<02:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.085     0.4608     0.1148        333        640:  82%|████████▏ | 664/809 [13:00<02:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.087     0.4601      0.115        360        640:  82%|████████▏ | 666/809 [13:02<02:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.09     0.4593     0.1151        345        640:  83%|████████▎ | 668/809 [13:04<02:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.091      0.459     0.1152        278        640:  83%|████████▎ | 669/809 [13:06<02:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.092     0.4584     0.1153        192        640:  83%|████████▎ | 671/809 [13:08<02:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.095     0.4576     0.1155        217        640:  83%|████████▎ | 673/809 [13:10<02:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.097      0.457     0.1156        157        640:  83%|████████▎ | 675/809 [13:13<02:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.098     0.4567     0.1156        359        640:  84%|████████▎ | 676/809 [13:14<02:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G        1.1     0.4561     0.1157        290        640:  84%|████████▍ | 678/809 [13:16<02:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.102     0.4556     0.1158        155        640:  84%|████████▍ | 680/809 [13:19<02:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.103     0.4552     0.1161        383        640:  84%|████████▍ | 681/809 [13:20<02:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.106     0.4544     0.1162        252        640:  84%|████████▍ | 683/809 [13:22<02:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.108     0.4538     0.1164        498        640:  85%|████████▍ | 685/809 [13:25<02:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.108     0.4536     0.1164        184        640:  85%|████████▍ | 686/809 [13:26<02:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.11     0.4532     0.1164        193        640:  85%|████████▌ | 688/809 [13:28<02:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.11     0.4533     0.1165        243        640:  85%|████████▌ | 690/809 [13:30<02:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.109     0.4534     0.1164        178        640:  86%|████████▌ | 692/809 [13:33<02:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.11     0.4532     0.1164        277        640:  86%|████████▌ | 693/809 [13:34<02:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.11     0.4529     0.1165        366        640:  86%|████████▌ | 695/809 [13:36<02:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.111     0.4527     0.1165        108        640:  86%|████████▌ | 697/809 [13:39<02:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.111     0.4525     0.1165        128        640:  86%|████████▋ | 699/809 [13:41<02:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.112     0.4523     0.1166        322        640:  87%|████████▋ | 700/809 [13:42<02:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.112     0.4521     0.1166        182        640:  87%|████████▋ | 702/809 [13:44<02:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.113     0.4517     0.1167        242        640:  87%|████████▋ | 704/809 [13:47<02:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.113     0.4514     0.1167        170        640:  87%|████████▋ | 705/809 [13:48<02:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.115     0.4508     0.1169        429        640:  87%|████████▋ | 707/809 [13:50<02:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.116     0.4504     0.1169        319        640:  88%|████████▊ | 709/809 [13:53<02:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.116     0.4504     0.1169         97        640:  88%|████████▊ | 710/809 [13:54<02:00,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.117     0.4501     0.1171        145        640:  88%|████████▊ | 712/809 [13:56<01:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.118     0.4497     0.1173        450        640:  88%|████████▊ | 714/809 [13:59<01:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.118     0.4497     0.1173        202        640:  88%|████████▊ | 715/809 [14:00<01:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.119     0.4493     0.1174        359        640:  89%|████████▊ | 717/809 [14:02<01:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4492     0.1175        268        640:  89%|████████▉ | 719/809 [14:05<01:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4493     0.1176        183        640:  89%|████████▉ | 720/809 [14:06<01:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.121      0.449     0.1178        281        640:  89%|████████▉ | 722/809 [14:08<01:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.121      0.449     0.1178        358        640:  89%|████████▉ | 724/809 [14:11<01:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.121      0.449     0.1178        350        640:  90%|████████▉ | 726/809 [14:13<01:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.121      0.449     0.1177        180        640:  90%|████████▉ | 727/809 [14:14<01:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.121     0.4489     0.1179        273        640:  90%|█████████ | 729/809 [14:16<01:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.122     0.4486      0.118        349        640:  90%|█████████ | 731/809 [14:19<01:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.122     0.4487     0.1179        167        640:  90%|█████████ | 732/809 [14:20<01:28,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.122     0.4488     0.1179        211        640:  91%|█████████ | 734/809 [14:22<01:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.122     0.4487     0.1179        196        640:  91%|█████████ | 736/809 [14:25<01:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.122     0.4487     0.1179        204        640:  91%|█████████ | 737/809 [14:26<01:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.123     0.4483     0.1181        271        640:  91%|█████████▏| 739/809 [14:28<01:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.124     0.4481     0.1181        269        640:  92%|█████████▏| 741/809 [14:30<01:18,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.124     0.4478     0.1183        239        640:  92%|█████████▏| 743/809 [14:33<01:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.124     0.4478     0.1183        288        640:  92%|█████████▏| 744/809 [14:34<01:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.125     0.4476     0.1183        470        640:  92%|█████████▏| 746/809 [14:36<01:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.125     0.4475     0.1185        336        640:  92%|█████████▏| 748/809 [14:39<01:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.125     0.4476     0.1184        106        640:  93%|█████████▎| 749/809 [14:40<01:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.124     0.4477     0.1184        215        640:  93%|█████████▎| 751/809 [14:42<01:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.125     0.4475     0.1184        424        640:  93%|█████████▎| 753/809 [14:45<01:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.125     0.4476     0.1184        380        640:  93%|█████████▎| 755/809 [14:47<01:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.125     0.4476     0.1185        293        640:  93%|█████████▎| 756/809 [14:48<01:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.125     0.4476     0.1185        277        640:  94%|█████████▎| 758/809 [14:50<00:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.124     0.4478     0.1184         84        640:  94%|█████████▍| 760/809 [14:53<00:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.124     0.4478     0.1184        208        640:  94%|█████████▍| 761/809 [14:54<00:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.124     0.4479     0.1184        173        640:  94%|█████████▍| 763/809 [14:56<00:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.123     0.4482     0.1183        145        640:  95%|█████████▍| 765/809 [14:58<00:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.124     0.4479     0.1184        190        640:  95%|█████████▍| 767/809 [15:01<00:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.124      0.448     0.1184        139        640:  95%|█████████▍| 768/809 [15:02<00:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.123     0.4482     0.1183        241        640:  95%|█████████▌| 770/809 [15:04<00:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.123     0.4483     0.1183        212        640:  95%|█████████▌| 772/809 [15:07<00:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.122     0.4485     0.1182        154        640:  96%|█████████▌| 774/809 [15:09<00:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.122     0.4486     0.1182        100        640:  96%|█████████▌| 775/809 [15:10<00:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.122     0.4486     0.1182        252        640:  96%|█████████▌| 777/809 [15:13<00:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.121     0.4486     0.1182        308        640:  96%|█████████▋| 779/809 [15:15<00:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.122     0.4485     0.1183        362        640:  96%|█████████▋| 780/809 [15:16<00:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.121     0.4486     0.1183        171        640:  97%|█████████▋| 782/809 [15:18<00:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.121     0.4486     0.1183        128        640:  97%|█████████▋| 784/809 [15:21<00:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.121     0.4485     0.1183        256        640:  97%|█████████▋| 786/809 [15:23<00:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4485     0.1183        364        640:  97%|█████████▋| 787/809 [15:24<00:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4485     0.1183        293        640:  98%|█████████▊| 789/809 [15:27<00:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4487     0.1183        239        640:  98%|█████████▊| 791/809 [15:29<00:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4486     0.1183        201        640:  98%|█████████▊| 793/809 [15:31<00:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4485     0.1184        240        640:  98%|█████████▊| 795/809 [15:33<00:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4487     0.1184        222        640:  98%|█████████▊| 796/809 [15:35<00:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4486     0.1184        279        640:  99%|█████████▊| 798/809 [15:37<00:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.119     0.4487     0.1184        525        640:  99%|█████████▉| 800/809 [15:39<00:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4486     0.1185        462        640:  99%|█████████▉| 801/809 [15:41<00:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4486     0.1185        302        640:  99%|█████████▉| 803/809 [15:43<00:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G       1.12     0.4485     0.1187        159        640: 100%|█████████▉| 805/809 [15:45<00:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.121     0.4483     0.1189        344        640: 100%|█████████▉| 807/809 [15:48<00:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     79/300      14.1G      1.121     0.4483     0.1189        455        640: 100%|█████████▉| 808/809 [15:49<00:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  10%|█         | 7/69 [00:01<00:14,  4.18it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  22%|██▏       | 15/69 [00:03<00:13,  4.08it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  33%|███▎      | 23/69 [00:05<00:11,  4.04it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  45%|████▍     | 31/69 [00:07<00:09,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  57%|█████▋    | 39/69 [00:09<00:07,  4.08it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  68%|██████▊   | 47/69 [00:11<00:06,  3.54it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  80%|███████▉  | 55/69 [00:13<00:03,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  91%|███████���█▏| 63/69 [00:15<00:01,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                   all        548      38759      0.529      0.363      0.368      0.206
-     80/300      13.8G      1.322     0.3668     0.1546        328        640:   0%|          | 0/809 [00:01<?, ?it/s]ning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.161     0.4108     0.1187        212        640:   0%|          | 3/809 [00:03<15:53,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.114     0.4264     0.1164        280        640:   0%|          | 4/809 [00:04<15:37,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.102       0.43     0.1171        257        640:   1%|          | 6/809 [00:07<15:37,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.067     0.4604     0.1109        152        640:   1%|          | 8/809 [00:09<15:38,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.066     0.4563     0.1105        320        640:   1%|          | 9/809 [00:10<15:38,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.073     0.4595     0.1113        292        640:   1%|▏         | 11/809 [00:12<15:26,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.106     0.4474     0.1116        335        640:   2%|▏         | 13/809 [00:15<15:32,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4437     0.1109        332        640:   2%|▏         | 15/809 [00:17<15:43,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.116     0.4482     0.1107        229        640:   2%|▏         | 16/809 [00:18<15:49,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4485     0.1103        269        640:   2%|▏         | 18/809 [00:21<15:56,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.12     0.4496     0.1092        255        640:   2%|▏         | 20/809 [00:23<15:41,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.115       0.45     0.1083        214        640:   3%|▎         | 21/809 [00:24<15:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.124       0.45      0.114        358        640:   3%|▎         | 23/809 [00:27<15:25,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4504     0.1145        266        640:   3%|▎         | 25/809 [00:29<15:24,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.134       0.45     0.1138        215        640:   3%|▎         | 26/809 [00:30<15:28,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.129     0.4507     0.1118        307        640:   3%|▎         | 28/809 [00:33<15:33,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.118     0.4517     0.1102        124        640:   4%|▎         | 30/809 [00:35<15:48,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4489     0.1104        159        640:   4%|▍         | 31/809 [00:36<15:28,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.16     0.4398     0.1209        355        640:   4%|▍         | 33/809 [00:39<15:43,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.194     0.4293     0.1269        381        640:   4%|▍         | 35/809 [00:41<15:21,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.206     0.4268     0.1291        130        640:   4%|▍         | 36/809 [00:42<15:11,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.214     0.4255     0.1279        232        640:   5%|▍         | 38/809 [00:45<15:21,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.221      0.423     0.1275        295        640:   5%|▍         | 40/809 [00:47<14:56,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.224     0.4216     0.1267        248        640:   5%|▌         | 42/809 [00:49<14:51,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.218     0.4224     0.1258        171        640:   5%|▌         | 43/809 [00:50<14:52,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.213     0.4245     0.1245        164        640:   6%|▌         | 45/809 [00:53<15:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.214     0.4233     0.1235        265        640:   6%|▌         | 47/809 [00:55<14:57,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.21     0.4245     0.1234        218        640:   6%|▌         | 48/809 [00:56<14:47,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.217     0.4232     0.1255        395        640:   6%|▌         | 50/809 [00:59<14:38,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.227       0.42     0.1259        294        640:   6%|▋         | 52/809 [01:01<14:37,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.227      0.421     0.1256        373        640:   7%|▋         | 53/809 [01:02<15:05,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.22     0.4223     0.1248        220        640:   7%|▋         | 55/809 [01:05<15:03,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.206     0.4272     0.1247        101        640:   7%|▋         | 57/809 [01:07<14:43,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.204     0.4288     0.1255        219        640:   7%|▋         | 59/809 [01:09<14:42,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G        1.2     0.4303     0.1251        123        640:   7%|▋         | 60/809 [01:10<14:32,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.197     0.4311     0.1249        284        640:   8%|▊         | 62/809 [01:13<14:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.194     0.4321     0.1244        227        640:   8%|▊         | 64/809 [01:15<14:25,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.193     0.4328     0.1241        148        640:   8%|▊         | 65/809 [01:16<14:20,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.191     0.4332     0.1237        212        640:   8%|▊         | 67/809 [01:18<14:16,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.19     0.4329     0.1231        249        640:   9%|▊         | 69/809 [01:21<14:13,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.188     0.4337     0.1222        228        640:   9%|▉         | 71/809 [01:23<14:25,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.19     0.4329     0.1219        184        640:   9%|▉         | 72/809 [01:24<14:19,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.188     0.4335      0.121        251        640:   9%|▉         | 74/809 [01:27<14:06,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.188     0.4334     0.1202        164        640:   9%|▉         | 76/809 [01:29<14:04,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.187     0.4346     0.1201        121        640:  10%|▉         | 78/809 [01:31<14:17,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.186     0.4353     0.1201        258        640:  10%|▉         | 79/809 [01:32<14:10,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.188     0.4356     0.1199        121        640:  10%|█         | 81/809 [01:35<14:12,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.183     0.4386     0.1195         99        640:  10%|█         | 83/809 [01:37<13:59,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.178     0.4405     0.1193         69        640:  10%|█         | 84/809 [01:38<13:59,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.179     0.4402     0.1196        237        640:  11%|█         | 86/809 [01:41<14:06,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.177     0.4418     0.1193        151        640:  11%|█         | 88/809 [01:43<13:50,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.173     0.4444     0.1189         61        640:  11%|█         | 90/809 [01:45<13:48,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.17     0.4455     0.1186        232        640:  11%|█         | 91/809 [01:46<13:43,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.17     0.4455     0.1183        161        640:  11%|█▏        | 93/809 [01:49<13:48,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.165     0.4474     0.1178        116        640:  12%|█▏        | 95/809 [01:51<13:48,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.166     0.4467     0.1174        160        640:  12%|█▏        | 97/809 [01:53<13:51,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.167     0.4467     0.1175        191        640:  12%|█▏        | 98/809 [01:55<14:07,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.168     0.4463     0.1171        243        640:  12%|█▏        | 100/809 [01:57<14:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.163      0.448     0.1167        204        640:  13%|█▎        | 102/809 [01:59<13:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.159     0.4489     0.1165        135        640:  13%|█▎        | 103/809 [02:00<13:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156     0.4494      0.116        207        640:  13%|█▎        | 105/809 [02:03<13:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.151     0.4504     0.1158         51        640:  13%|█▎        | 107/809 [02:05<13:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152       0.45     0.1164        348        640:  13%|█▎        | 108/809 [02:06<13:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155      0.449     0.1192        165        640:  14%|█▎        | 110/809 [02:09<14:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.154     0.4489     0.1189        193        640:  14%|█▎        | 111/809 [02:10<15:15,  1.31s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.154     0.4482     0.1184        182        640:  14%|█▍        | 113/809 [02:13<14:12,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155     0.4472     0.1199        343        640:  14%|█▍        | 115/809 [02:15<13:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156     0.4465     0.1199        225        640:  14%|█▍        | 116/809 [02:16<13:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.153     0.4472     0.1198        396        640:  15%|█▍        | 118/809 [02:19<14:18,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156     0.4465     0.1209        375        640:  15%|█▍        | 120/809 [02:21<14:21,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155     0.4468     0.1207        281        640:  15%|█▍        | 121/809 [02:23<14:07,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.157      0.446     0.1209        293        640:  15%|█▌        | 123/809 [02:25<13:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155     0.4466     0.1208        275        640:  15%|█▌        | 125/809 [02:27<13:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4474     0.1206        141        640:  16%|█▌        | 126/809 [02:28<13:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155     0.4468     0.1209        204        640:  16%|█▌        | 128/809 [02:31<13:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.161     0.4449     0.1219        384        640:  16%|█▌        | 130/809 [02:33<13:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.164     0.4444     0.1237        255        640:  16%|█▋        | 132/809 [02:35<13:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.169      0.443     0.1248        265        640:  16%|█▋        | 133/809 [02:37<13:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.171     0.4425     0.1248        258        640:  17%|█▋        | 135/809 [02:39<13:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.172      0.442     0.1247        207        640:  17%|█▋        | 137/809 [02:41<12:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.172     0.4422     0.1246        239        640:  17%|█▋        | 139/809 [02:44<13:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.172     0.4421     0.1245        228        640:  17%|█▋        | 141/809 [02:46<13:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.173     0.4417     0.1247        233        640:  18%|█▊        | 142/809 [02:47<13:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.177     0.4408     0.1258        242        640:  18%|█▊        | 144/809 [02:50<13:36,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.177     0.4405     0.1258        210        640:  18%|█▊        | 145/809 [02:51<13:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.175     0.4404     0.1255        271        640:  18%|█▊        | 147/809 [02:53<13:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.176     0.4395     0.1253        306        640:  18%|█▊        | 149/809 [02:56<13:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.174     0.4398      0.125         90        640:  19%|█▊        | 151/809 [02:58<13:19,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.172     0.4403     0.1248        122        640:  19%|█▉        | 152/809 [02:59<13:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.172     0.4402     0.1252        221        640:  19%|█▉        | 154/809 [03:02<12:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.174     0.4389     0.1255        451        640:  19%|█▉        | 156/809 [03:04<12:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.175     0.4389     0.1256        257        640:  19%|█▉        | 157/809 [03:05<12:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.171     0.4396     0.1252        138        640:  20%|█▉        | 159/809 [03:07<12:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.17     0.4397     0.1248        201        640:  20%|█▉        | 161/809 [03:10<12:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.17     0.4395     0.1247        235        640:  20%|██        | 163/809 [03:12<12:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.17     0.4393     0.1246        348        640:  20%|██        | 164/809 [03:13<12:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.169     0.4392     0.1245        170        640:  21%|██        | 166/809 [03:16<12:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.168     0.4396     0.1248        176        640:  21%|██        | 168/809 [03:18<12:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.168     0.4394     0.1248        442        640:  21%|██        | 169/809 [03:19<12:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.166     0.4402     0.1243         97        640:  21%|██        | 171/809 [03:22<12:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.165     0.4403     0.1241        239        640:  21%|██▏       | 173/809 [03:24<12:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.165     0.4401     0.1239        232        640:  22%|██▏       | 175/809 [03:26<12:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.164     0.4404     0.1236        223        640:  22%|██▏       | 176/809 [03:27<12:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.163     0.4407     0.1234        334        640:  22%|██▏       | 178/809 [03:30<12:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.162     0.4408     0.1234        100        640:  22%|██▏       | 180/809 [03:32<12:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.161     0.4409     0.1233        145        640:  22%|██▏       | 181/809 [03:33<12:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.16     0.4414     0.1231        372        640:  23%|██▎       | 183/809 [03:36<12:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.159     0.4415     0.1229        209        640:  23%|██▎       | 185/809 [03:38<12:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.157     0.4421     0.1228        205        640:  23%|██▎       | 187/809 [03:40<12:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156     0.4423     0.1225        167        640:  23%|██▎       | 188/809 [03:41<12:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155     0.4424     0.1221        221        640:  23%|██▎       | 190/809 [03:44<12:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155     0.4418      0.122         70        640:  24%|██▎       | 192/809 [03:46<11:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156     0.4413     0.1221        338        640:  24%|██▍       | 193/809 [03:47<12:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.16     0.4403     0.1227        314        640:  24%|██▍       | 195/809 [03:49<11:48,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.158     0.4401     0.1225        118        640:  24%|██▍       | 197/809 [03:52<11:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.157     0.4402     0.1224        240        640:  25%|██▍       | 199/809 [03:54<11:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155      0.441     0.1221         76        640:  25%|██▍       | 200/809 [03:55<11:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.154     0.4411     0.1219        152        640:  25%|██▍       | 202/809 [03:58<11:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4422     0.1219        185        640:  25%|██▌       | 204/809 [04:00<11:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.151     0.4427     0.1217        100        640:  25%|██▌       | 205/809 [04:01<11:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.149     0.4436     0.1216        206        640:  26%|██▌       | 207/809 [04:04<11:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.149     0.4433     0.1218        415        640:  26%|██▌       | 209/809 [04:06<11:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.15      0.443     0.1219        389        640:  26%|██▌       | 211/809 [04:08<11:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.149     0.4435     0.1218        217        640:  26%|██▌       | 212/809 [04:09<11:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.147      0.444     0.1215        131        640:  26%|██▋       | 214/809 [04:12<11:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.151     0.4427     0.1222        386        640:  27%|██▋       | 216/809 [04:14<11:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.151     0.4425     0.1221        244        640:  27%|██▋       | 217/809 [04:15<11:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4418     0.1221        480        640:  27%|██▋       | 219/809 [04:18<11:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.151     0.4421      0.122        238        640:  27%|██▋       | 221/809 [04:20<11:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152      0.442     0.1221        309        640:  27%|██▋       | 222/809 [04:21<11:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.15     0.4426     0.1218        348        640:  28%|██▊       | 224/809 [04:24<11:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.148     0.4429     0.1215        263        640:  28%|██▊       | 226/809 [04:26<11:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4434     0.1213        178        640:  28%|██▊       | 227/809 [04:27<11:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.141     0.4443      0.121         72        640:  28%|██▊       | 229/809 [04:30<11:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.141     0.4443      0.121        269        640:  29%|██▊       | 231/809 [04:32<11:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.141     0.4443      0.121        242        640:  29%|██▊       | 232/809 [04:33<11:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.141     0.4441     0.1208        262        640:  29%|██▉       | 234/809 [04:36<11:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.143     0.4436     0.1207        237        640:  29%|██▉       | 236/809 [04:38<11:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4442     0.1204        104        640:  29%|██▉       | 238/809 [04:40<11:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4442     0.1204        303        640:  30%|██▉       | 239/809 [04:41<11:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139     0.4445     0.1201        214        640:  30%|██▉       | 241/809 [04:44<10:53,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139     0.4441     0.1199        247        640:  30%|███       | 243/809 [04:46<11:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4438     0.1205        174        640:  30%|███       | 245/809 [04:48<10:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139     0.4441     0.1203        238        640:  30%|███       | 246/809 [04:49<10:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.138     0.4443     0.1201        166        640:  31%|███       | 248/809 [04:52<10:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.136     0.4449     0.1199        108        640:  31%|███       | 250/809 [04:54<10:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.137     0.4446     0.1199        252        640:  31%|███       | 251/809 [04:55<10:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.135      0.445     0.1197        309        640:  31%|███▏      | 253/809 [04:58<10:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.136     0.4447     0.1201        347        640:  32%|███▏      | 255/809 [05:00<10:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.136     0.4446     0.1204        406        640:  32%|███▏      | 257/809 [05:02<10:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.136     0.4447     0.1203        232        640:  32%|███▏      | 258/809 [05:03<10:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.135     0.4449     0.1201        268        640:  32%|███▏      | 260/809 [05:06<10:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.137     0.4446     0.1206        143        640:  32%|███▏      | 262/809 [05:08<11:10,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.137     0.4445     0.1206        350        640:  33%|███▎      | 263/809 [05:10<11:27,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4436     0.1209        567        640:  33%|███▎      | 265/809 [05:12<10:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4434      0.121        218        640:  33%|███▎      | 267/809 [05:14<10:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.142     0.4428     0.1216        436        640:  33%|███▎      | 268/809 [05:15<10:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.143     0.4425     0.1221        213        640:  33%|███▎      | 270/809 [05:18<10:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.141     0.4432     0.1218        121        640:  34%|███▎      | 272/809 [05:20<10:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.141      0.443     0.1218        279        640:  34%|███▎      | 273/809 [05:21<10:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139     0.4436     0.1216        407        640:  34%|███▍      | 275/809 [05:24<10:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.138     0.4435     0.1214        156        640:  34%|███▍      | 277/809 [05:26<10:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.138     0.4433     0.1213        113        640:  34%|███▍      | 279/809 [05:28<10:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.138     0.4433     0.1214        195        640:  35%|███▍      | 280/809 [05:30<10:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139      0.443     0.1217        374        640:  35%|███▍      | 282/809 [05:32<10:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139     0.4429     0.1218        406        640:  35%|███▍      | 283/809 [05:33<10:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4425     0.1218        293        640:  35%|███▌      | 285/809 [05:36<10:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139     0.4425     0.1217        342        640:  35%|███▌      | 287/809 [05:38<10:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139     0.4424     0.1219        269        640:  36%|███▌      | 289/809 [05:41<10:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.141     0.4422      0.123        489        640:  36%|███▌      | 291/809 [05:43<10:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.141     0.4423     0.1229        344        640:  36%|███▌      | 292/809 [05:44<10:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4422      0.123        278        640:  36%|███▋      | 294/809 [05:46<10:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.142     0.4417     0.1232        391        640:  37%|███▋      | 296/809 [05:49<10:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.142     0.4416     0.1231        180        640:  37%|███▋      | 297/809 [05:50<09:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.143     0.4414     0.1234        284        640:  37%|███▋      | 299/809 [05:52<09:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.145     0.4411     0.1236        494        640:  37%|███▋      | 301/809 [05:55<10:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.145      0.441     0.1236        312        640:  37%|███▋      | 302/809 [05:56<10:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.145      0.441     0.1235        269        640:  38%|███▊      | 304/809 [05:58<09:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4407     0.1234        279        640:  38%|███▊      | 306/809 [06:01<09:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4409     0.1233        122        640:  38%|███▊      | 308/809 [06:03<09:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4407     0.1233        184        640:  38%|███▊      | 309/809 [06:04<09:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146      0.441     0.1232        259        640:  38%|███▊      | 311/809 [06:07<10:18,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.147     0.4406     0.1234        614        640:  39%|███▊      | 312/809 [06:08<10:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4407     0.1236        603        640:  39%|███▉      | 314/809 [06:10<10:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.148     0.4404     0.1238        601        640:  39%|███▉      | 316/809 [06:13<09:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.149       0.44     0.1238        317        640:  39%|███▉      | 318/809 [06:15<09:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.15     0.4399     0.1239        249        640:  39%|███▉      | 319/809 [06:16<09:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.151     0.4395     0.1239        252        640:  40%|███▉      | 321/809 [06:19<09:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.154     0.4385      0.124        198        640:  40%|███▉      | 323/809 [06:21<09:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155     0.4384      0.124        261        640:  40%|████      | 324/809 [06:22<09:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.159     0.4374     0.1246        610        640:  40%|████      | 326/809 [06:24<09:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.158      0.438     0.1244        152        640:  41%|████      | 328/809 [06:27<09:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.158      0.438     0.1243        339        640:  41%|████      | 330/809 [06:29<09:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.158      0.438     0.1242        195        640:  41%|████      | 331/809 [06:30<09:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.157     0.4383     0.1242        120        640:  41%|████      | 333/809 [06:33<09:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.157     0.4383      0.124        121        640:  41%|████▏     | 335/809 [06:35<09:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.158      0.438     0.1241        540        640:  42%|████▏     | 336/809 [06:36<09:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.158     0.4382     0.1241        189        640:  42%|████▏     | 338/809 [06:39<09:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.159     0.4379     0.1242        538        640:  42%|████▏     | 340/809 [06:41<09:39,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.159      0.438     0.1241        238        640:  42%|████▏     | 341/809 [06:42<09:33,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.16     0.4375     0.1242        261        640:  42%|████▏     | 343/809 [06:45<09:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.161      0.437      0.124        238        640:  43%|████▎     | 345/809 [06:47<09:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.162      0.437     0.1241        133        640:  43%|████▎     | 346/809 [06:48<08:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.163     0.4364     0.1246        115        640:  43%|████▎     | 348/809 [06:50<08:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.165     0.4359     0.1256        352        640:  43%|████▎     | 350/809 [06:53<09:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.165     0.4358     0.1255        462        640:  44%|████▎     | 352/809 [06:55<08:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.164      0.436     0.1254        208        640:  44%|████▎     | 353/809 [06:57<09:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.162     0.4367     0.1252        222        640:  44%|████▍     | 355/809 [06:59<08:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.162     0.4365      0.125        363        640:  44%|████▍     | 357/809 [07:01<08:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.161     0.4367     0.1248        294        640:  44%|████▍     | 358/809 [07:02<08:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.159     0.4372     0.1246        102        640:  44%|████▍     | 360/809 [07:05<08:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.16     0.4369     0.1245        195        640:  45%|████▍     | 362/809 [07:07<08:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.159     0.4369     0.1243        324        640:  45%|████▍     | 364/809 [07:09<08:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.159     0.4368     0.1243        357        640:  45%|████▌     | 365/809 [07:10<08:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.159     0.4369     0.1241        258        640:  45%|████▌     | 367/809 [07:13<08:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.158     0.4373      0.124        286        640:  46%|████▌     | 369/809 [07:15<08:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.157     0.4375     0.1239         99        640:  46%|████▌     | 370/809 [07:16<08:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156     0.4377     0.1237        240        640:  46%|████▌     | 372/809 [07:19<08:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156     0.4377     0.1237        215        640:  46%|████▌     | 374/809 [07:21<08:49,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156     0.4377     0.1237        294        640:  46%|████▋     | 375/809 [07:22<08:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.158     0.4372     0.1238        308        640:  47%|████▋     | 377/809 [07:25<08:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.159     0.4371     0.1238        251        640:  47%|████▋     | 379/809 [07:27<08:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.159      0.437     0.1237        284        640:  47%|████▋     | 380/809 [07:28<08:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.158     0.4371     0.1236        148        640:  47%|████▋     | 382/809 [07:31<08:37,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.157     0.4374     0.1235        270        640:  47%|████▋     | 384/809 [07:33<08:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156     0.4377     0.1235         49        640:  48%|████▊     | 385/809 [07:34<08:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156     0.4378     0.1235        302        640:  48%|████▊     | 387/809 [07:37<08:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.157     0.4377     0.1241        520        640:  48%|████▊     | 389/809 [07:39<08:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156      0.438      0.124        149        640:  48%|████▊     | 390/809 [07:40<08:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.156      0.438      0.124        228        640:  48%|████▊     | 392/809 [07:43<08:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155     0.4379     0.1239        167        640:  49%|████▊     | 394/809 [07:45<08:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155     0.4379     0.1239        165        640:  49%|████▉     | 395/809 [07:46<08:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155     0.4381     0.1237        136        640:  49%|████▉     | 397/809 [07:49<08:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.155      0.438     0.1238        431        640:  49%|████▉     | 399/809 [07:51<08:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.154     0.4382     0.1236        158        640:  49%|████▉     | 400/809 [07:52<08:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.153     0.4386     0.1235        124        640:  50%|████▉     | 402/809 [07:55<07:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4388     0.1234        272        640:  50%|████▉     | 404/809 [07:57<07:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4388     0.1232        263        640:  50%|█████     | 406/809 [07:59<07:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.153     0.4387     0.1234        247        640:  50%|█████     | 407/809 [08:00<07:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4387     0.1232        230        640:  51%|█████     | 409/809 [08:03<08:11,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4385     0.1231        245        640:  51%|█████     | 411/809 [08:05<07:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4385     0.1232        230        640:  51%|█████     | 412/809 [08:06<07:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4388     0.1231        234        640:  51%|█████     | 414/809 [08:09<07:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4389     0.1231         92        640:  51%|█████▏    | 416/809 [08:11<07:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152      0.439      0.123        202        640:  52%|█████▏    | 417/809 [08:12<07:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.151     0.4394     0.1229        142        640:  52%|█████▏    | 419/809 [08:15<07:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.151     0.4395     0.1228        198        640:  52%|█████▏    | 421/809 [08:17<07:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4393     0.1229        506        640:  52%|█████▏    | 422/809 [08:18<07:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.152     0.4393     0.1228        282        640:  52%|█████▏    | 424/809 [08:21<07:50,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.15     0.4398     0.1227        166        640:  53%|█████▎    | 426/809 [08:23<07:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.149     0.4398     0.1226        388        640:  53%|█████▎    | 428/809 [08:25<07:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.147     0.4403     0.1223        171        640:  53%|█████▎    | 430/809 [08:28<07:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.147     0.4403     0.1222        335        640:  53%|█████▎    | 431/809 [08:29<07:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4405     0.1222        156        640:  54%|█████▎    | 433/809 [08:31<07:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4404     0.1222        226        640:  54%|█████▍    | 435/809 [08:34<07:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.147     0.4403     0.1224        400        640:  54%|█████▍    | 437/809 [08:36<07:09,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.147     0.4403     0.1223        264        640:  54%|█████▍    | 438/809 [08:37<07:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4404     0.1223        195        640:  54%|█████▍    | 440/809 [08:39<07:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4406     0.1221        186        640:  55%|█████▍    | 442/809 [08:42<07:00,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4405     0.1221        264        640:  55%|█████▍    | 444/809 [08:44<07:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.145     0.4408     0.1221        127        640:  55%|█████▌    | 445/809 [08:45<07:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4408     0.1222        221        640:  55%|█████▌    | 447/809 [08:48<07:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.145     0.4412     0.1221        131        640:  55%|█████▌    | 448/809 [08:49<07:27,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.145     0.4411      0.122        543        640:  56%|█████▌    | 450/809 [08:51<07:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.144     0.4413      0.122        236        640:  56%|���████▌    | 452/809 [08:54<07:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.145     0.4412     0.1221        427        640:  56%|█████▌    | 454/809 [08:56<07:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.144     0.4412     0.1219        189        640:  56%|█████▌    | 455/809 [08:57<07:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4408     0.1225        540        640:  56%|█████▋    | 457/809 [09:00<06:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.147     0.4407     0.1227        359        640:  57%|█████▋    | 459/809 [09:02<06:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4407     0.1226        207        640:  57%|█████▋    | 460/809 [09:03<07:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.145      0.441     0.1224        230        640:  57%|█████▋    | 462/809 [09:06<07:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4408     0.1225        210        640:  57%|█████▋    | 464/809 [09:08<06:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.147     0.4405     0.1226        364        640:  57%|█████▋    | 465/809 [09:09<06:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146     0.4409     0.1224        155        640:  58%|█████▊    | 467/809 [09:12<06:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.145     0.4411     0.1222        166        640:  58%|█████▊    | 469/809 [09:14<06:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.146      0.441     0.1222        189        640:  58%|█████▊    | 470/809 [09:15<06:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.145     0.4412     0.1221        136        640:  58%|█████▊    | 472/809 [09:17<06:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.144     0.4413     0.1219        292        640:  59%|█████▊    | 474/809 [09:20<06:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.143     0.4416     0.1217        209        640:  59%|█████▉    | 476/809 [09:22<06:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.143     0.4417     0.1216        261        640:  59%|█████▉    | 477/809 [09:23<06:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.142     0.4419     0.1215        137        640:  59%|█████▉    | 479/809 [09:26<06:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.141     0.4423     0.1213        151        640:  59%|█████▉    | 481/809 [09:28<06:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4425     0.1212        193        640:  60%|█████▉    | 482/809 [09:29<06:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.141     0.4425     0.1213        583        640:  60%|█████▉    | 484/809 [09:31<06:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4426     0.1212        196        640:  60%|██████    | 486/809 [09:34<06:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139     0.4426     0.1211        241        640:  60%|██████    | 487/809 [09:35<06:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.138      0.443     0.1209        209        640:  60%|██████    | 489/809 [09:37<06:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139     0.4428     0.1208        490        640:  61%|██████    | 491/809 [09:40<06:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4426      0.121        261        640:  61%|██████    | 493/809 [09:42<06:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4426      0.121        290        640:  61%|██████    | 494/809 [09:43<06:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.14     0.4427     0.1208        361        640:  61%|██████▏   | 496/809 [09:46<06:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139      0.443     0.1208        249        640:  62%|██████▏   | 498/809 [09:48<06:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.139     0.4431     0.1207        203        640:  62%|██████▏   | 499/809 [09:49<06:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.138     0.4431     0.1206        342        640:  62%|██████▏   | 501/809 [09:51<05:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.138     0.4433     0.1206        329        640:  62%|██████▏   | 503/809 [09:54<06:18,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.137     0.4434     0.1205        276        640:  62%|██████▏   | 504/809 [09:55<06:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.137     0.4436     0.1205        290        640:  63%|██████▎   | 506/809 [09:57<05:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.136     0.4438     0.1205        346        640:  63%|██████▎   | 508/809 [10:00<05:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.135     0.4442     0.1205        362        640:  63%|██████▎   | 510/809 [10:02<05:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.135     0.4444     0.1204        201        640:  63%|██████▎   | 511/809 [10:03<05:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.134     0.4444     0.1204        315        640:  63%|██████▎   | 513/809 [10:06<05:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.133     0.4448     0.1203        237        640:  64%|██████▎   | 515/809 [10:08<05:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.134     0.4448     0.1203        295        640:  64%|██████▍   | 517/809 [10:10<05:37,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.133     0.4451     0.1202        146        640:  64%|██████▍   | 518/809 [10:11<05:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.133      0.445     0.1202        352        640:  64%|██████▍   | 520/809 [10:14<05:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.132     0.4453     0.1201        160        640:  65%|██████▍   | 522/809 [10:16<05:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4455       0.12         79        640:  65%|██████▍   | 523/809 [10:17<05:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4457     0.1199        323        640:  65%|██████▍   | 525/809 [10:20<05:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.129     0.4462     0.1198        166        640:  65%|██████▌   | 527/809 [10:22<05:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.129     0.4461     0.1197        378        640:  65%|██████▌   | 528/809 [10:23<05:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4458       0.12        405        640:  66%|██████▌   | 530/809 [10:26<05:45,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4457       0.12        178        640:  66%|██████▌   | 532/809 [10:28<05:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4457     0.1199        257        640:  66%|██████▌   | 533/809 [10:29<05:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4457     0.1199        185        640:  66%|██████▌   | 535/809 [10:32<05:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.132     0.4455     0.1198        244        640:  66%|██████▋   | 537/809 [10:34<05:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4456     0.1198        379        640:  67%|██████▋   | 539/809 [10:36<05:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4458     0.1197        202        640:  67%|██████▋   | 540/809 [10:38<05:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4458     0.1197        356        640:  67%|██████▋   | 542/809 [10:40<05:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4462     0.1196        101        640:  67%|██████▋   | 544/809 [10:42<05:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4464     0.1196        125        640:  67%|██████▋   | 545/809 [10:43<05:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.132     0.4461     0.1204        290        640:  68%|██████▊   | 547/809 [10:46<05:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4464     0.1202        285        640:  68%|██████▊   | 549/809 [10:48<05:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4465     0.1202        196        640:  68%|██████▊   | 550/809 [10:49<05:15,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4462     0.1203        272        640:  68%|██████▊   | 552/809 [10:52<05:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4463     0.1202        140        640:  68%|██████▊   | 554/809 [10:54<04:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4464     0.1201        280        640:  69%|██████▊   | 556/809 [10:56<04:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4462     0.1202        344        640:  69%|██████▉   | 557/809 [10:58<04:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4464     0.1201        205        640:  69%|██████▉   | 559/809 [11:00<04:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4465       0.12        271        640:  69%|██████▉   | 561/809 [11:02<04:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4465     0.1199        264        640:  69%|██████▉   | 562/809 [11:03<04:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4466     0.1198        151        640:  70%|██████▉   | 564/809 [11:06<04:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4467     0.1198        375        640:  70%|██████▉   | 566/809 [11:08<04:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4466     0.1197        273        640:  70%|███████   | 568/809 [11:10<04:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4467     0.1197        161        640:  70%|███████   | 569/809 [11:11<04:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4467     0.1197        104        640:  71%|███████   | 571/809 [11:14<04:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4467     0.1197        253        640:  71%|███████   | 573/809 [11:16<04:31,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.129     0.4468     0.1196        206        640:  71%|███████   | 575/809 [11:18<04:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.129     0.4467     0.1196        349        640:  71%|███████   | 576/809 [11:20<04:27,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4467     0.1195        313        640:  71%|███████▏  | 578/809 [11:22<04:25,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.129     0.4467     0.1195        207        640:  72%|███████▏  | 580/809 [11:24<04:23,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.129     0.4468     0.1194        333        640:  72%|███████▏  | 582/809 [11:26<04:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.128     0.4469     0.1193        149        640:  72%|███████▏  | 583/809 [11:28<04:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4465     0.1198        446        640:  72%|███████▏  | 585/809 [11:30<04:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4463     0.1202        551        640:  73%|███████▎  | 587/809 [11:33<04:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4462     0.1201        344        640:  73%|███████▎  | 588/809 [11:34<04:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4461     0.1201        173        640:  73%|███████▎  | 590/809 [11:36<04:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4459     0.1201        573        640:  73%|███████▎  | 592/809 [11:38<04:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4459     0.1201        250        640:  73%|███████▎  | 594/809 [11:41<04:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4459       0.12        305        640:  74%|███████▎  | 596/809 [11:43<04:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4458       0.12        220        640:  74%|███████▍  | 597/809 [11:44<04:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.132     0.4455     0.1205        485        640:  74%|███████▍  | 599/809 [11:47<04:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.133     0.4454     0.1205        220        640:  74%|███████▍  | 601/809 [11:49<04:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.132     0.4457     0.1205        141        640:  74%|███████▍  | 602/809 [11:50<04:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.132     0.4457     0.1207        167        640:  75%|███████▍  | 604/809 [11:53<03:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4459     0.1206        290        640:  75%|███████▍  | 606/809 [11:55<03:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4462     0.1205        304        640:  75%|███████▌  | 608/809 [11:57<03:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131      0.446     0.1205        272        640:  75%|███████▌  | 609/809 [11:58<03:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131      0.446     0.1206        169        640:  76%|███████▌  | 611/809 [12:01<03:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.131     0.4462     0.1206        302        640:  76%|███████▌  | 613/809 [12:03<03:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G       1.13     0.4464     0.1205        140        640:  76%|███████▌  | 614/809 [12:04<03:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.129     0.4468     0.1204        229        640:  76%|███████▌  | 616/809 [12:07<03:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.129     0.4469     0.1203        236        640:  76%|███████▋  | 618/809 [12:09<03:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.128     0.4472     0.1202        152        640:  77%|███████▋  | 619/809 [12:10<03:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4476     0.1201         95        640:  77%|███████▋  | 621/809 [12:13<03:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4477     0.1201        193        640:  77%|███████▋  | 623/809 [12:15<03:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4477     0.1201        354        640:  77%|███████▋  | 624/809 [12:16<03:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4477       0.12        331        640:  77%|███████▋  | 626/809 [12:18<03:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4478       0.12        245        640:  78%|███████▊  | 628/809 [12:21<03:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4479       0.12        311        640:  78%|███████▊  | 629/809 [12:22<03:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4477     0.1201        195        640:  78%|███████▊  | 631/809 [12:24<03:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4477     0.1202        200        640:  78%|███████▊  | 633/809 [12:27<03:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126      0.448     0.1201        184        640:  78%|███████▊  | 635/809 [12:29<03:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4482       0.12        117        640:  79%|███████▊  | 636/809 [12:30<03:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4481     0.1203        438        640:  79%|███████▉  | 638/809 [12:33<03:27,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4477     0.1204        170        640:  79%|███████▉  | 640/809 [12:35<03:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4477     0.1204        221        640:  79%|███████▉  | 641/809 [12:36<03:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.128     0.4474     0.1205        360        640:  79%|███████▉  | 643/809 [12:38<03:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.128     0.4473     0.1204        265        640:  80%|███████▉  | 645/809 [12:41<03:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.128     0.4473     0.1203        215        640:  80%|███████▉  | 647/809 [12:43<03:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.128     0.4473     0.1203        271        640:  80%|████████  | 648/809 [12:44<03:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.128     0.4473     0.1203        224        640:  80%|████████  | 650/809 [12:47<03:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4473     0.1201        277        640:  81%|████████  | 652/809 [12:49<03:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4475     0.1201        195        640:  81%|████████  | 653/809 [12:50<03:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4475       0.12        271        640:  81%|████████  | 655/809 [12:53<03:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4476     0.1199        293        640:  81%|████████  | 657/809 [12:55<02:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4476     0.1199        303        640:  81%|████████▏ | 658/809 [12:56<02:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4479     0.1197        173        640:  82%|████████▏ | 660/809 [12:59<02:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4481     0.1196        197        640:  82%|████████▏ | 662/809 [13:01<03:01,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4483     0.1195        171        640:  82%|████████▏ | 663/809 [13:02<02:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4482     0.1195        440        640:  82%|████████▏ | 665/809 [13:05<02:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4481     0.1199        253        640:  82%|████████▏ | 667/809 [13:07<02:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4481     0.1199        348        640:  83%|████████▎ | 668/809 [13:08<02:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123      0.448     0.1198        308        640:  83%|████████▎ | 670/809 [13:11<02:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4482     0.1197        170        640:  83%|████████▎ | 672/809 [13:13<02:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4482     0.1201        204        640:  83%|████████▎ | 674/809 [13:15<02:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4482       0.12        312        640:  83%|████████▎ | 675/809 [13:16<02:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4484       0.12        319        640:  84%|████████▎ | 677/809 [13:19<02:31,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4485       0.12        285        640:  84%|████████▍ | 679/809 [13:21<02:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4488     0.1199        256        640:  84%|████████▍ | 681/809 [13:23<02:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4488     0.1198        431        640:  84%|████████▍ | 682/809 [13:25<02:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4489     0.1197        157        640:  85%|████████▍ | 684/809 [13:27<02:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122      0.449     0.1199        123        640:  85%|████████▍ | 686/809 [13:29<02:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4488     0.1199        320        640:  85%|████████▍ | 687/809 [13:30<02:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122      0.449     0.1199        136        640:  85%|████████▌ | 689/809 [13:33<02:18,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4491     0.1199        268        640:  85%|████████▌ | 691/809 [13:35<02:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4489       0.12        347        640:  86%|████████▌ | 692/809 [13:36<02:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4491     0.1199        205        640:  86%|████████▌ | 694/809 [13:39<02:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4489       0.12        387        640:  86%|████████▌ | 696/809 [13:41<02:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.124     0.4487     0.1201        486        640:  86%|████████▋ | 698/809 [13:43<02:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.124     0.4487       0.12        150        640:  86%|████████▋ | 699/809 [13:44<02:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.124     0.4485     0.1202        193        640:  87%|████████▋ | 701/809 [13:47<02:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4485     0.1201        332        640:  87%|████████▋ | 703/809 [13:49<02:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4485     0.1201        471        640:  87%|████████▋ | 704/809 [13:50<02:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.124     0.4487       0.12        205        640:  87%|████████▋ | 706/809 [13:53<01:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.124     0.4489     0.1199        251        640:  88%|████████▊ | 708/809 [13:55<01:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4491     0.1198        358        640:  88%|████████▊ | 710/809 [13:57<01:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4492     0.1197        167        640:  88%|████████▊ | 711/809 [13:58<01:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4494     0.1197        230        640:  88%|████████▊ | 713/809 [14:01<01:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4494     0.1196        291        640:  88%|████████▊ | 715/809 [14:03<01:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4494     0.1197        369        640:  89%|████████▊ | 717/809 [14:05<01:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4495     0.1197        268        640:  89%|████████▉ | 718/809 [14:07<01:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4497     0.1195        181        640:  89%|████████▉ | 720/809 [14:09<01:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4497     0.1195        174        640:  89%|████████▉ | 722/809 [14:11<01:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4498     0.1194        220        640:  89%|████████▉ | 723/809 [14:13<01:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122     0.4499     0.1194        234        640:  90%|████████▉ | 725/809 [14:15<01:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4499     0.1193        362        640:  90%|████████▉ | 727/809 [14:17<01:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.122       0.45     0.1193        173        640:  90%|████████▉ | 728/809 [14:18<01:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4498     0.1197        602        640:  90%|█████████ | 730/809 [14:21<01:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4498     0.1197        148        640:  90%|█████████ | 732/809 [14:23<01:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4498     0.1196        256        640:  91%|█████████ | 734/809 [14:25<01:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4498     0.1196        286        640:  91%|█████████ | 735/809 [14:27<01:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4501     0.1195        111        640:  91%|█████████ | 737/809 [14:29<01:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123       0.45     0.1194        271        640:  91%|█████████▏| 739/809 [14:31<01:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4499     0.1195        335        640:  92%|█████████▏| 741/809 [14:34<01:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.123     0.4499     0.1195        194        640:  92%|█████████▏| 743/809 [14:36<01:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.124     0.4496     0.1198        561        640:  92%|█████████▏| 744/809 [14:37<01:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.124     0.4497     0.1198        315        640:  92%|█████████▏| 746/809 [14:39<01:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4495       0.12        120        640:  92%|█████████▏| 748/809 [14:42<01:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4494       0.12        409        640:  93%|█████████▎| 749/809 [14:43<01:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4494       0.12        331        640:  93%|█████████▎| 751/809 [14:45<01:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4494     0.1201        538        640:  93%|█████████▎| 753/809 [14:48<01:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4496       0.12        179        640:  93%|█████████▎| 754/809 [14:49<01:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4495       0.12        256        640:  93%|█████████▎| 756/809 [14:51<01:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4495       0.12        117        640:  94%|█████████▎| 758/809 [14:54<00:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4493     0.1202        142        640:  94%|█████████▍| 760/809 [14:56<00:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4491     0.1202        367        640:  94%|█████████▍| 761/809 [14:57<00:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4489     0.1203        352        640:  94%|█████████▍| 763/809 [15:00<00:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4487     0.1203        337        640:  95%|█████████▍| 765/809 [15:02<00:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4487     0.1203        187        640:  95%|█████████▍| 766/809 [15:03<00:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4487     0.1203        313        640:  95%|█████████▍| 768/809 [15:05<00:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4488     0.1202        201        640:  95%|█████████▌| 770/809 [15:08<00:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4489     0.1201        308        640:  95%|█████████▌| 772/809 [15:10<00:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4488     0.1202        529        640:  96%|█████████▌| 773/809 [15:11<00:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4489     0.1201        289        640:  96%|█████████▌| 775/809 [15:14<00:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4488       0.12        289        640:  96%|█████████▌| 777/809 [15:16<00:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4487       0.12        304        640:  96%|█████████▌| 778/809 [15:17<00:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4489     0.1199        182        640:  96%|█████████▋| 780/809 [15:19<00:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4489     0.1198        280        640:  97%|█████████▋| 782/809 [15:22<00:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125      0.449     0.1198        133        640:  97%|█████████▋| 784/809 [15:24<00:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4489     0.1197        211        640:  97%|█████████▋| 785/809 [15:25<00:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125      0.449     0.1197        215        640:  97%|█████████▋| 787/809 [15:28<00:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125      0.449     0.1197        490        640:  98%|█████████▊| 789/809 [15:30<00:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125      0.449     0.1196        229        640:  98%|█████████▊| 791/809 [15:32<00:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125      0.449     0.1196        269        640:  98%|█████████▊| 792/809 [15:33<00:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4488     0.1197        494        640:  98%|█████████▊| 794/809 [15:36<00:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4488     0.1198        130        640:  98%|█████████▊| 796/809 [15:38<00:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.125     0.4487     0.1198        289        640:  99%|█████████▊| 797/809 [15:39<00:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4487     0.1197        330        640:  99%|█████████▉| 799/809 [15:42<00:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4485     0.1197        259        640:  99%|█████████▉| 801/809 [15:44<00:09,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.126     0.4483     0.1197        247        640:  99%|█████████▉| 802/809 [15:45<00:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4482     0.1197        288        640:  99%|█████████▉| 804/809 [15:48<00:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127      0.448     0.1197        278        640: 100%|█████████▉| 806/809 [15:50<00:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     80/300      13.9G      1.127     0.4481     0.1197        209        640: 100%|█████████▉| 807/809 [15:51<00:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   3%|▎         | 2/69 [00:00<00:16,  4.09it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  16%|█▌        | 11/69 [00:02<00:14,  4.14it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  28%|██▊       | 19/69 [00:04<00:12,  4.12it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  39%|███▉      | 27/69 [00:06<00:10,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  51%|█████     | 35/69 [00:08<00:08,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  62%|██████▏   | 43/69 [00:10<00:06,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  74%|███████▍  | 51/69 [00:12<00:04,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  86%|████████▌ | 59/69 [00:14<00:02,  4.08it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:16<00:00,  4.10it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:16<00:00,  4.10it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:16<00:00,  4.07it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:16<00:00,  4.07it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:16<00:00,  4.07it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.183     0.3967     0.1281        257        640:   0%|          | 2/809 [00:02<15:53,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.142     0.4167     0.1184        294        640:   0%|          | 4/809 [00:04<15:28,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.167     0.4212      0.119        231        640:   1%|          | 5/809 [00:05<15:36,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.102     0.4538     0.1065         99        640:   1%|          | 7/809 [00:08<15:47,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.075     0.4583     0.1063        262        640:   1%|          | 9/809 [00:10<15:27,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.073     0.4609     0.1052        232        640:   1%|          | 10/809 [00:11<15:17,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.114     0.4483     0.1074        151        640:   1%|▏         | 12/809 [00:13<15:27,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.085     0.4595     0.1032        156        640:   2%|▏         | 14/809 [00:16<15:18,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4499     0.1082        148        640:   2%|▏         | 16/809 [00:18<15:13,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.125     0.4468     0.1084        247        640:   2%|▏         | 17/809 [00:19<15:19,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.134     0.4447     0.1085        293        640:   2%|▏         | 19/809 [00:22<15:13,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.137     0.4413     0.1119        176        640:   3%|���         | 21/809 [00:24<15:19,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.16     0.4323     0.1132        355        640:   3%|▎         | 22/809 [00:26<15:10,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.171     0.4289     0.1131        255        640:   3%|▎         | 24/809 [00:27<15:07,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.177     0.4265     0.1112         91        640:   3%|▎         | 26/809 [00:30<15:07,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.182     0.4235      0.111        231        640:   3%|▎         | 28/809 [00:32<15:21,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.183      0.423     0.1104        203        640:   4%|▎         | 29/809 [00:33<15:33,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.175     0.4261     0.1084        223        640:   4%|▍         | 31/809 [00:36<15:27,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.148     0.4317      0.106        135        640:   4%|▍         | 33/809 [00:38<15:06,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.141      0.433      0.105        228        640:   4%|▍         | 34/809 [00:39<15:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.136     0.4363     0.1054        149        640:   4%|▍         | 36/809 [00:41<14:57,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.139     0.4352     0.1054        204        640:   5%|▍         | 38/809 [00:44<15:08,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.155     0.4303     0.1174        503        640:   5%|▍         | 40/809 [00:46<15:14,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.161      0.428     0.1177        320        640:   5%|▌         | 41/809 [00:47<14:59,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.16     0.4299     0.1179        343        640:   5%|▌         | 43/809 [00:50<14:52,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.161     0.4292     0.1169        177        640:   6%|▌         | 45/809 [00:52<14:52,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.164     0.4277      0.117        262        640:   6%|▌         | 46/809 [00:53<14:59,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.175     0.4251     0.1264        250        640:   6%|▌         | 48/809 [00:56<15:00,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.178     0.4231     0.1276        402        640:   6%|▌         | 50/809 [00:58<14:56,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.173     0.4245      0.127         95        640:   6%|▋         | 51/809 [00:59<15:00,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.188     0.4204     0.1292        406        640:   7%|▋         | 53/809 [01:02<15:01,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.188     0.4201     0.1279        265        640:   7%|▋         | 55/809 [01:04<14:57,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.183     0.4225     0.1274        207        640:   7%|▋         | 57/809 [01:06<14:38,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.187     0.4222     0.1292        425        640:   7%|▋         | 58/809 [01:07<14:29,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.182     0.4241     0.1279        199        640:   7%|▋         | 60/809 [01:10<14:24,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.177     0.4261     0.1267        190        640:   8%|▊         | 62/809 [01:12<14:47,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.173     0.4279     0.1264        144        640:   8%|▊         | 63/809 [01:13<14:34,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.177     0.4272     0.1292        149        640:   8%|▊         | 65/809 [01:16<14:41,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.174     0.4287     0.1283        232        640:   8%|▊         | 67/809 [01:18<14:25,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.171     0.4305     0.1285        138        640:   8%|▊         | 68/809 [01:20<14:20,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.171     0.4309     0.1285        303        640:   9%|▊         | 70/809 [01:22<14:22,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.163     0.4335     0.1273         87        640:   9%|▉         | 72/809 [01:24<14:10,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.166     0.4331     0.1281        297        640:   9%|▉         | 74/809 [01:26<14:05,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.161     0.4346     0.1273        159        640:   9%|▉         | 75/809 [01:27<14:02,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.166     0.4335     0.1277        589        640:  10%|▉         | 77/809 [01:30<14:47,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.17     0.4325      0.128        261        640:  10%|▉         | 79/809 [01:32<14:27,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.171     0.4325     0.1282        309        640:  10%|█         | 81/809 [01:34<14:05,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.17      0.432     0.1282        265        640:  10%|█         | 82/809 [01:36<14:12,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.168     0.4325     0.1283        293        640:  10%|█         | 84/809 [01:38<14:44,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.171     0.4316     0.1284        293        640:  11%|█         | 85/809 [01:39<14:58,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.177     0.4299     0.1298        331        640:  11%|█         | 87/809 [01:42<14:41,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.181     0.4292     0.1303        411        640:  11%|█         | 89/809 [01:44<14:33,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.179       0.43       0.13        219        640:  11%|█         | 90/809 [01:46<14:48,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.181     0.4287     0.1297        244        640:  11%|█▏        | 92/809 [01:48<14:16,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.18     0.4286     0.1294        251        640:  12%|█▏        | 94/809 [01:50<14:11,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.182     0.4284     0.1295        464        640:  12%|█▏        | 95/809 [01:51<14:09,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.177     0.4303     0.1288        182        640:  12%|█▏        | 97/809 [01:54<13:46,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.173     0.4314     0.1283        227        640:  12%|█▏        | 99/809 [01:56<13:38,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.166     0.4347     0.1278        140        640:  12%|█▏        | 101/809 [01:58<13:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.158      0.437     0.1271        276        640:  13%|█▎        | 103/809 [02:01<13:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.158     0.4374     0.1283        434        640:  13%|█▎        | 105/809 [02:03<13:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.156     0.4379     0.1279        188        640:  13%|█▎        | 106/809 [02:04<13:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.153     0.4382      0.128        144        640:  13%|█▎        | 108/809 [02:06<13:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.157     0.4374     0.1307        248        640:  14%|█▎        | 110/809 [02:09<14:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.159     0.4366     0.1306        322        640:  14%|█▎        | 111/809 [02:10<13:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.163     0.4355     0.1324        424        640:  14%|█▍        | 113/809 [02:13<13:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.166     0.4345     0.1324        314        640:  14%|█▍        | 115/809 [02:15<13:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.165     0.4346      0.132        187        640:  14%|█▍        | 116/809 [02:16<13:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.163     0.4349     0.1315        214        640:  15%|█▍        | 118/809 [02:18<13:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.162     0.4356     0.1325        153        640:  15%|█▍        | 120/809 [02:21<13:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.161     0.4355     0.1319        309        640:  15%|█▌        | 122/809 [02:23<13:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.162     0.4352      0.132        313        640:  15%|█▌        | 123/809 [02:24<13:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.163     0.4353      0.132        437        640:  15%|█▌        | 125/809 [02:27<13:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.164      0.435     0.1325        296        640:  16%|█▌        | 127/809 [02:29<13:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.163      0.435     0.1321        227        640:  16%|█▌        | 128/809 [02:30<13:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.168     0.4334     0.1344        384        640:  16%|█▌        | 130/809 [02:33<13:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.172     0.4317     0.1354        291        640:  16%|█▋        | 132/809 [02:35<13:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.177     0.4302     0.1368        662        640:  16%|█▋        | 133/809 [02:36<14:00,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.179     0.4296     0.1368        383        640:  17%|█▋        | 135/809 [02:39<13:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.176     0.4303     0.1362        170        640:  17%|█▋        | 137/809 [02:41<13:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.176       0.43     0.1359        257        640:  17%|█▋        | 138/809 [02:42<13:49,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.177     0.4299     0.1358        320        640:  17%|█▋        | 140/809 [02:45<13:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.176     0.4301     0.1355        250        640:  18%|█▊        | 142/809 [02:47<13:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.177     0.4298     0.1354        319        640:  18%|█▊        | 143/809 [02:48<13:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.178     0.4298     0.1351        221        640:  18%|█▊        | 145/809 [02:51<12:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.176     0.4308     0.1346        215        640:  18%|█▊        | 147/809 [02:53<12:43,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.178     0.4303     0.1346        318        640:  18%|█▊        | 148/809 [02:54<12:42,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.177     0.4306     0.1342        252        640:  19%|█▊        | 150/809 [02:56<12:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.174     0.4316     0.1337        158        640:  19%|█▉        | 152/809 [02:59<13:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.175     0.4313     0.1334        250        640:  19%|█▉        | 154/809 [03:01<12:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.173     0.4322      0.133        147        640:  19%|█▉        | 155/809 [03:02<12:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.172     0.4326     0.1326        234        640:  19%|█▉        | 157/809 [03:05<12:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.174     0.4319     0.1324        332        640:  20%|█▉        | 159/809 [03:07<12:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.174      0.432     0.1323        294        640:  20%|█▉        | 160/809 [03:08<12:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.172     0.4326      0.132        280        640:  20%|██        | 162/809 [03:11<12:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.17     0.4331     0.1316        272        640:  20%|██        | 164/809 [03:13<12:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.167      0.434     0.1312        200        640:  20%|██        | 165/809 [03:14<12:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.169     0.4332     0.1311        415        640:  21%|██        | 167/809 [03:16<12:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.173     0.4322     0.1312        412        640:  21%|██        | 169/809 [03:19<12:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.173     0.4321     0.1309        187        640:  21%|██        | 171/809 [03:21<12:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.173      0.432     0.1309        164        640:  21%|██▏       | 172/809 [03:22<12:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.171     0.4325     0.1305        206        640:  22%|██▏       | 174/809 [03:25<12:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.169     0.4324     0.1302        248        640:  22%|██▏       | 176/809 [03:27<12:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.167     0.4329     0.1299        165        640:  22%|██▏       | 177/809 [03:28<12:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.165     0.4334       0.13        503        640:  22%|██▏       | 179/809 [03:31<12:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.163     0.4344     0.1295        210        640:  22%|██▏       | 181/809 [03:33<12:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.162     0.4348     0.1294        382        640:  23%|██▎       | 183/809 [03:35<12:02,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.162     0.4349     0.1292        305        640:  23%|██▎       | 184/809 [03:36<12:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.162     0.4352     0.1289        206        640:  23%|██▎       | 186/809 [03:39<11:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.161     0.4361     0.1286         93        640:  23%|██▎       | 188/809 [03:41<11:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.162     0.4361     0.1287        255        640:  23%|██▎       | 190/809 [03:43<12:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.163     0.4359     0.1287        542        640:  24%|██▎       | 191/809 [03:45<12:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.161     0.4364     0.1283        235        640:  24%|██▍       | 193/809 [03:47<11:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.16     0.4366     0.1279        331        640:  24%|██▍       | 195/809 [03:49<11:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.16     0.4367     0.1278        224        640:  24%|██▍       | 196/809 [03:50<11:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.161     0.4364     0.1279        252        640:  24%|██▍       | 198/809 [03:53<12:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.163     0.4362     0.1279        280        640:  25%|██▍       | 200/809 [03:55<11:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.165     0.4353     0.1281        364        640:  25%|██▍       | 202/809 [03:57<11:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.167     0.4349     0.1285        226        640:  25%|██▌       | 203/809 [03:58<11:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.171     0.4339     0.1286        303        640:  25%|██▌       | 205/809 [04:01<11:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.171     0.4337     0.1287        206        640:  26%|██▌       | 207/809 [04:03<11:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.17     0.4338     0.1285        247        640:  26%|██▌       | 208/809 [04:04<11:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.17     0.4338     0.1282        292        640:  26%|██▌       | 210/809 [04:07<11:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.167     0.4344     0.1278        314        640:  26%|██▌       | 212/809 [04:09<11:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.166      0.435      0.128        358        640:  26%|██▋       | 214/809 [04:11<11:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.167     0.4348     0.1279        410        640:  27%|██▋       | 215/809 [04:12<11:25,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.165      0.435     0.1278        304        640:  27%|██▋       | 217/809 [04:15<11:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.162     0.4356     0.1277        227        640:  27%|██▋       | 219/809 [04:17<11:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.162     0.4358     0.1275        287        640:  27%|██▋       | 220/809 [04:18<11:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.161     0.4359     0.1274        109        640:  27%|██▋       | 222/809 [04:21<11:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.162     0.4358     0.1271        282        640:  28%|██▊       | 224/809 [04:23<11:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.16      0.436     0.1267        266        640:  28%|██▊       | 226/809 [04:25<11:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.16     0.4362     0.1265        324        640:  28%|██▊       | 227/809 [04:26<11:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.158     0.4368     0.1263        163        640:  28%|██▊       | 229/809 [04:29<11:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.158     0.4366     0.1262        262        640:  29%|██▊       | 231/809 [04:31<11:06,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.158     0.4371     0.1263        392        640:  29%|██▉       | 233/809 [04:33<11:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.156     0.4373     0.1261        216        640:  29%|██▉       | 234/809 [04:35<11:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.158     0.4374     0.1261        210        640:  29%|██▉       | 236/809 [04:37<10:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.158     0.4373     0.1263        116        640:  29%|██▉       | 238/809 [04:39<11:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.157     0.4374     0.1261        369        640:  30%|██▉       | 239/809 [04:40<11:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.157     0.4378     0.1259        225        640:  30%|██▉       | 241/809 [04:43<11:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.157      0.438     0.1259        424        640:  30%|███       | 243/809 [04:45<10:52,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.156     0.4381     0.1257        265        640:  30%|███       | 244/809 [04:46<10:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.156     0.4383     0.1256        254        640:  30%|███       | 246/809 [04:49<10:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.155     0.4386     0.1255        246        640:  31%|███       | 248/809 [04:51<10:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.153     0.4393     0.1252        179        640:  31%|███       | 250/809 [04:53<10:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.151     0.4398      0.125        143        640:  31%|███       | 251/809 [04:54<10:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.151     0.4398     0.1257        331        640:  31%|███▏      | 253/809 [04:57<10:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.152     0.4391     0.1255        124        640:  32%|███▏      | 255/809 [04:59<10:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.157     0.4378     0.1259        375        640:  32%|███▏      | 257/809 [05:01<10:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.158     0.4374     0.1262        276        640:  32%|███▏      | 258/809 [05:02<10:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.159     0.4371     0.1261        185        640:  32%|███▏      | 260/809 [05:05<10:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.159     0.4369      0.126        164        640:  32%|███▏      | 262/809 [05:07<10:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.159     0.4369      0.126        249        640:  33%|███▎      | 263/809 [05:08<10:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.159     0.4367     0.1258        456        640:  33%|███▎      | 265/809 [05:11<10:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.16     0.4365     0.1257        169        640:  33%|███▎      | 267/809 [05:13<10:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.161     0.4362     0.1259        323        640:  33%|█��█▎      | 269/809 [05:15<10:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.161     0.4369     0.1258        115        640:  33%|███▎      | 270/809 [05:17<10:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.157     0.4381     0.1253        232        640:  34%|███▎      | 273/809 [05:20<10:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.159     0.4376      0.126        310        640:  34%|███▍      | 274/809 [05:21<10:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.16     0.4371     0.1261        505        640:  34%|███▍      | 276/809 [05:24<10:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.159     0.4376     0.1264        235        640:  34%|███▍      | 278/809 [05:26<10:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.159     0.4377     0.1263        333        640:  34%|███▍      | 279/809 [05:27<10:09,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.157     0.4382     0.1261         87        640:  35%|███▍      | 281/809 [05:29<10:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.156     0.4387     0.1259        224        640:  35%|███▍      | 283/809 [05:32<10:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.156     0.4386     0.1259        255        640:  35%|███▌      | 285/809 [05:34<10:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.155     0.4391     0.1258        130        640:  35%|███▌      | 286/809 [05:35<10:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.154     0.4392     0.1258        236        640:  36%|███▌      | 288/809 [05:38<10:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.154      0.439     0.1257        220        640:  36%|███▌      | 290/809 [05:40<09:59,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.154     0.4394     0.1256        187        640:  36%|███▌      | 292/809 [05:42<09:56,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.153     0.4395     0.1255        294        640:  36%|███▌      | 293/809 [05:43<10:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.152     0.4402     0.1252        285        640:  36%|███▋      | 295/809 [05:46<09:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.152     0.4403     0.1252        156        640:  37%|███▋      | 297/809 [05:48<09:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.152     0.4402     0.1252        342        640:  37%|███▋      | 298/809 [05:49<09:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.151     0.4404      0.125        331        640:  37%|███▋      | 300/809 [05:52<10:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.149     0.4411     0.1247         84        640:  37%|███▋      | 302/809 [05:54<10:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.15      0.441     0.1246        240        640:  37%|███▋      | 303/809 [05:55<09:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.152     0.4403     0.1246        297        640:  38%|███▊      | 305/809 [05:57<09:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.152     0.4401     0.1246        144        640:  38%|███▊      | 307/809 [06:00<09:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.152     0.4402     0.1245        246        640:  38%|███▊      | 309/809 [06:02<09:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.152     0.4402     0.1244        236        640:  38%|███▊      | 310/809 [06:03<09:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.149     0.4409     0.1239        179        640:  39%|███▊      | 312/809 [06:05<09:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.147     0.4415     0.1237        183        640:  39%|███▉      | 314/809 [06:08<09:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.146     0.4413     0.1239        190        640:  39%|███▉      | 316/809 [06:10<09:25,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.145     0.4414     0.1238        127        640:  39%|███▉      | 317/809 [06:11<09:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.144     0.4418     0.1236        220        640:  39%|███▉      | 319/809 [06:14<09:22,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.146     0.4413     0.1237        178        640:  40%|███▉      | 321/809 [06:16<09:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.148     0.4407     0.1238        155        640:  40%|███▉      | 323/809 [06:18<09:20,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.15     0.4403     0.1241        215        640:  40%|████      | 324/809 [06:19<09:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.148     0.4409      0.124         94        640:  40%|████      | 326/809 [06:22<09:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.146     0.4415     0.1237        108        640:  41%|████      | 328/809 [06:24<09:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.146     0.4415     0.1237        251        640:  41%|████      | 329/809 [06:25<09:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.145     0.4416     0.1235        182        640:  41%|████      | 331/809 [06:28<09:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.146     0.4414     0.1234        271        640:  41%|████      | 333/809 [06:30<09:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.146     0.4417     0.1234         90        640:  41%|████▏     | 334/809 [06:31<09:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.149     0.4409     0.1241        238        640:  42%|████▏     | 336/809 [06:34<09:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.15     0.4407     0.1241        238        640:  42%|████▏     | 338/809 [06:36<09:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.151     0.4404     0.1242        380        640:  42%|████▏     | 340/809 [06:38<09:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.152     0.4401     0.1242        356        640:  42%|████▏     | 341/809 [06:39<09:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.151     0.4404     0.1241        225        640:  42%|████▏     | 343/809 [06:42<09:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.15     0.4406     0.1239        316        640:  43%|████▎     | 345/809 [06:44<09:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.149     0.4406     0.1237        281        640:  43%|████▎     | 346/809 [06:45<09:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.149     0.4407     0.1235        138        640:  43%|████▎     | 348/809 [06:48<08:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.15     0.4404     0.1235        215        640:  43%|████▎     | 350/809 [06:50<08:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.151     0.4399     0.1236         95        640:  44%|████▎     | 352/809 [06:52<08:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.151     0.4399     0.1236        310        640:  44%|████▎     | 353/809 [06:53<08:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.151       0.44     0.1238        404        640:  44%|████▍     | 355/809 [06:56<09:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.15     0.4403     0.1237        230        640:  44%|████▍     | 357/809 [06:58<08:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.149     0.4404     0.1237        225        640:  44%|████▍     | 358/809 [06:59<08:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.148     0.4409     0.1234        236        640:  44%|████▍     | 360/809 [07:02<08:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.146     0.4414     0.1231        197        640:  45%|████▍     | 362/809 [07:04<08:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.146     0.4415     0.1235        467        640:  45%|████▍     | 364/809 [07:06<08:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.145      0.442     0.1234         98        640:  45%|████▌     | 365/809 [07:08<08:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.145     0.4419     0.1234        170        640:  45%|████▌     | 367/809 [07:10<08:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.143     0.4424     0.1231        202        640:  46%|████▌     | 369/809 [07:12<08:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.142     0.4425      0.123        201        640:  46%|████▌     | 370/809 [07:13<08:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.142     0.4425     0.1229        165        640:  46%|████▌     | 372/809 [07:16<08:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.141     0.4429     0.1227        235        640:  46%|████▌     | 374/809 [07:18<08:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.14     0.4431     0.1227        407        640:  46%|████▋     | 376/809 [07:20<08:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.139     0.4433     0.1228        205        640:  47%|████▋     | 377/809 [07:21<08:17,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.139     0.4434     0.1228        295        640:  47%|████▋     | 379/809 [07:24<08:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.139     0.4438     0.1227        115        640:  47%|████▋     | 381/809 [07:26<08:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.139     0.4436     0.1227        370        640:  47%|████▋     | 382/809 [07:27<08:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.139     0.4437     0.1227        400        640:  47%|████▋     | 384/809 [07:30<08:37,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.138      0.444     0.1226        370        640:  48%|████▊     | 386/809 [07:32<08:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.138     0.4442     0.1225         90        640:  48%|████▊     | 387/809 [07:33<08:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.138     0.4442     0.1228        102        640:  48%|████▊     | 389/809 [07:36<08:37,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.138     0.4439     0.1226        170        640:  48%|████▊     | 391/809 [07:38<08:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.138     0.4437     0.1225        385        640:  48%|████▊     | 392/809 [07:39<08:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.139     0.4434     0.1228        487        640:  49%|████▊     | 394/809 [07:42<08:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.138     0.4434     0.1227        296        640:  49%|████▉     | 396/809 [07:44<08:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.138     0.4433     0.1227        387        640:  49%|████▉     | 398/809 [07:46<08:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.138     0.4436     0.1225        133        640:  49%|████▉     | 399/809 [07:48<07:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.137     0.4439     0.1225        241        640:  50%|████▉     | 401/809 [07:50<07:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.137     0.4439     0.1226        323        640:  50%|████▉     | 403/809 [07:52<07:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.137     0.4439     0.1227        440        640:  50%|████▉     | 404/809 [07:53<07:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.137      0.444     0.1225        275        640:  50%|█████     | 406/809 [07:56<07:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.138     0.4436     0.1225        242        640:  50%|█████     | 408/809 [07:58<07:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.14      0.443     0.1228        457        640:  51%|█████     | 409/809 [07:59<08:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.141     0.4426     0.1227        233        640:  51%|█████     | 411/809 [08:02<07:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.143     0.4421     0.1228        270        640:  51%|█████     | 413/809 [08:04<07:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.143      0.442     0.1228        418        640:  51%|█████▏    | 415/809 [08:06<07:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.143     0.4419     0.1229        453        640:  52%|█████▏    | 417/809 [08:09<07:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.144     0.4417     0.1232        392        640:  52%|█████▏    | 418/809 [08:10<07:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.145     0.4414     0.1232        357        640:  52%|█████▏    | 420/809 [08:12<07:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.145     0.4415     0.1232        518        640:  52%|█████▏    | 422/809 [08:15<07:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.144     0.4416      0.123        369        640:  52%|█████▏    | 424/809 [08:17<07:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.144     0.4417      0.123        215        640:  53%|█████▎    | 425/809 [08:18<07:47,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.143     0.4419     0.1228        129        640:  53%|█████▎    | 427/809 [08:21<07:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.143     0.4417      0.123        300        640:  53%|█████▎    | 428/809 [08:22<07:43,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.143     0.4418     0.1228        223        640:  53%|█████▎    | 430/809 [08:24<07:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.142     0.4422     0.1225        329        640:  53%|█████▎    | 432/809 [08:27<07:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.141     0.4426     0.1223        141        640:  54%|█████▎    | 434/809 [08:29<07:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.141     0.4427     0.1222        268        640:  54%|█████▍    | 435/809 [08:30<07:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.142     0.4425     0.1225        562        640:  54%|█████▍    | 437/809 [08:33<07:28,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.141     0.4427     0.1224        115        640:  54%|█████▍    | 439/809 [08:35<07:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.141     0.4428     0.1223        305        640:  54%|█████▍    | 440/809 [08:36<07:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.14     0.4432     0.1221        359        640:  55%|█████▍    | 442/809 [08:38<07:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.14     0.4431     0.1222        415        640:  55%|█████▍    | 444/809 [08:41<07:35,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.14     0.4433     0.1221        168        640:  55%|█████▌    | 445/809 [08:42<07:33,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.139     0.4435      0.122        204        640:  55%|█████▌    | 447/809 [08:45<07:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.139     0.4437      0.122        279        640:  56%|█████▌    | 449/809 [08:47<07:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.139     0.4437     0.1219        221        640:  56%|█████▌    | 450/809 [08:48<07:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.138     0.4439     0.1222        384        640:  56%|█████▌    | 452/809 [08:50<06:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.137     0.4442      0.122        160        640:  56%|█████▌    | 454/809 [08:53<06:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.136     0.4446     0.1219        222        640:  56%|█████▋    | 456/809 [08:55<06:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.136     0.4446     0.1218        198        640:  56%|█████▋    | 457/809 [08:56<06:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.135     0.4446     0.1217        351        640:  57%|█████▋    | 459/809 [08:59<06:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.134     0.4447     0.1216        328        640:  57%|█████▋    | 461/809 [09:01<06:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.135     0.4445     0.1216        323        640:  57%|█████▋    | 462/809 [09:02<06:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.134     0.4446     0.1214        139        640:  57%|█████▋    | 464/809 [09:05<06:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.134     0.4446     0.1213        226        640:  58%|█████▊    | 466/809 [09:07<06:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.134     0.4445     0.1213        122        640:  58%|█████▊    | 467/809 [09:08<06:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.134     0.4443     0.1212        357        640:  58%|█████▊    | 469/809 [09:11<06:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.134     0.4443     0.1211        145        640:  58%|█████▊    | 471/809 [09:13<06:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.134     0.4444     0.1211         76        640:  58%|█████▊    | 472/809 [09:14<06:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.134     0.4443     0.1211        270        640:  59%|█████▊    | 474/809 [09:17<06:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.135     0.4442     0.1213        411        640:  59%|█████▉    | 476/809 [09:19<06:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.134     0.4443     0.1213        141        640:  59%|█████▉    | 477/809 [09:20<06:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.135     0.4444     0.1217        244        640:  59%|█████▉    | 479/809 [09:22<06:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.133     0.4446     0.1215        167        640:  59%|█████▉    | 481/809 [09:25<06:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.133     0.4446     0.1215        313        640:  60%|█████▉    | 483/809 [09:27<06:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.132     0.4448     0.1215        349        640:  60%|█████▉    | 484/809 [09:28<06:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.133     0.4448     0.1215        344        640:  60%|██████    | 486/809 [09:31<06:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4449     0.1213        335        640:  60%|██████    | 488/809 [09:33<06:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4451     0.1212        100        640:  60%|██████    | 489/809 [09:34<06:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4452     0.1211        157        640:  61%|██████    | 491/809 [09:37<06:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4451      0.121        470        640:  61%|██████    | 493/809 [09:39<06:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4451     0.1209        193        640:  61%|██████    | 495/809 [09:41<06:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4451     0.1209        337        640:  61%|██████▏   | 496/809 [09:42<06:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4452     0.1208        334        640:  62%|██████▏   | 498/809 [09:45<06:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4454     0.1206        226        640:  62%|██████▏   | 500/809 [09:47<05:56,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4454     0.1206        278        640:  62%|██████▏   | 502/809 [09:49<05:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4453     0.1206        343        640:  62%|██████▏   | 503/809 [09:50<05:53,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4453     0.1206        282        640:  62%|██████▏   | 505/809 [09:53<05:48,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129      0.445     0.1207        401        640:  63%|██████▎   | 507/809 [09:55<05:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129      0.445     0.1206        149        640:  63%|██████▎   | 508/809 [09:56<05:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4448     0.1207        362        640:  63%|██████▎   | 510/809 [09:59<05:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4447     0.1207        283        640:  63%|██████▎   | 512/809 [10:01<05:41,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4447     0.1207        350        640:  64%|██████▎   | 514/809 [10:03<05:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4446     0.1207        256        640:  64%|██████▎   | 515/809 [10:04<05:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4445     0.1211        436        640:  64%|██████▍   | 517/809 [10:07<05:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.132     0.4442     0.1214        337        640:  64%|██████▍   | 519/809 [10:09<05:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.132     0.4442     0.1215        403        640:  64%|██████▍   | 520/809 [10:10<05:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.132     0.4441     0.1216        294        640:  65%|██████▍   | 522/809 [10:13<05:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.132     0.4441     0.1216        242        640:  65%|██████▍   | 524/809 [10:15<05:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.132     0.4442     0.1215        220        640:  65%|██████▍   | 525/809 [10:16<05:46,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4443     0.1216        110        640:  65%|██████▌   | 527/809 [10:19<05:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.132     0.4441     0.1219        626        640:  65%|██████▌   | 529/809 [10:21<05:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.132     0.4443     0.1219        359        640:  66%|██████▌   | 531/809 [10:23<05:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.132     0.4443     0.1218        274        640:  66%|██████▌   | 532/809 [10:25<05:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4447     0.1217        146        640:  66%|██████▌   | 534/809 [10:27<05:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4444     0.1221        498        640:  66%|██████▋   | 536/809 [10:29<05:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4444      0.122        148        640:  66%|██████▋   | 537/809 [10:31<05:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4444      0.122        391        640:  67%|██████▋   | 539/809 [10:33<05:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4445     0.1219        174        640:  67%|██████▋   | 541/809 [10:35<05:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4445     0.1221        317        640:  67%|██████▋   | 542/809 [10:36<05:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4447      0.122        275        640:  67%|██████▋   | 544/809 [10:39<05:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4446      0.122        358        640:  67%|██████▋   | 546/809 [10:41<05:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4447     0.1218        261        640:  68%|██████▊   | 548/809 [10:43<05:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4449     0.1217        220        640:  68%|██████▊   | 549/809 [10:45<05:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4449     0.1217        338        640:  68%|██████▊   | 551/809 [10:47<05:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4447     0.1219        362        640:  68%|██████▊   | 553/809 [10:49<04:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129      0.445     0.1218        199        640:  68%|██████▊   | 554/809 [10:50<04:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4452     0.1217        441        640:  69%|██████▊   | 556/809 [10:53<04:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4449     0.1218        326        640:  69%|██████▉   | 558/809 [10:55<05:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129      0.445     0.1217        220        640:  69%|██████▉   | 559/809 [10:56<04:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4453     0.1217         91        640:  69%|██████▉   | 561/809 [10:59<04:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129      0.445     0.1218        434        640:  70%|██████▉   | 563/809 [11:01<04:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129      0.445     0.1219        139        640:  70%|██████▉   | 565/809 [11:04<04:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128      0.445     0.1218        229        640:  70%|███████   | 567/809 [11:06<04:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4451     0.1217        282        640:  70%|███████   | 568/809 [11:07<04:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4449     0.1215        208        640:  70%|███████   | 570/809 [11:10<04:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128      0.445     0.1214        331        640:  71%|███████   | 572/809 [11:12<04:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4448     0.1215        331        640:  71%|███████   | 573/809 [11:13<04:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.127     0.4451     0.1215        197        640:  71%|███████   | 575/809 [11:16<04:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4449     0.1219        478        640:  71%|███████▏  | 577/809 [11:18<04:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4448     0.1219        312        640:  71%|███████▏  | 578/809 [11:19<04:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.127      0.445     0.1218        273        640:  72%|███████▏  | 580/809 [11:22<04:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4449     0.1218        269        640:  72%|███████▏  | 582/809 [11:24<04:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4447     0.1219        446        640:  72%|███████▏  | 583/809 [11:25<04:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4445     0.1221        452        640:  72%|███████▏  | 585/809 [11:28<04:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4445     0.1221        151        640:  73%|███████▎  | 587/809 [11:30<04:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4442     0.1222        628        640:  73%|███████▎  | 588/809 [11:31<04:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131      0.444     0.1223        111        640:  73%|███████▎  | 590/809 [11:33<04:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4438     0.1223        168        640:  73%|███████▎  | 592/809 [11:36<04:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4438     0.1222        286        640:  73%|███████▎  | 594/809 [11:38<04:08,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4436     0.1222        380        640:  74%|███████▎  | 595/809 [11:39<04:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.131     0.4438     0.1221         94        640:  74%|███████▍  | 597/809 [11:41<04:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.13     0.4441      0.122         98        640:  74%|███████▍  | 599/809 [11:44<04:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4445     0.1219        244        640:  74%|███████▍  | 601/809 [11:46<04:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.128     0.4445     0.1218        201        640:  74%|███████▍  | 602/809 [11:47<03:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.129     0.4444     0.1218        288        640:  75%|███████▍  | 604/809 [11:50<03:55,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.127     0.4449     0.1217        134        640:  75%|███████▍  | 606/809 [11:52<03:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.126     0.4452     0.1216        346        640:  75%|███████▌  | 608/809 [11:54<03:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.126     0.4453     0.1216        263        640:  75%|███████▌  | 609/809 [11:55<03:48,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.126     0.4454     0.1215        108        640:  76%|███████▌  | 611/809 [11:58<03:48,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.125     0.4455     0.1215        270        640:  76%|███████▌  | 613/809 [12:00<03:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.125     0.4456     0.1214        161        640:  76%|███████▌  | 614/809 [12:01<03:44,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.125     0.4456     0.1216        211        640:  76%|███████▌  | 616/809 [12:03<03:42,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.125     0.4458     0.1215        210        640:  76%|███████▋  | 618/809 [12:06<03:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.124     0.4463     0.1213         74        640:  77%|███████▋  | 620/809 [12:08<03:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.124     0.4462     0.1213        402        640:  77%|███████▋  | 621/809 [12:09<03:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.124     0.4463     0.1212        272        640:  77%|███████▋  | 623/809 [12:12<03:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.123     0.4464     0.1211        145        640:  77%|███████▋  | 625/809 [12:14<03:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.122     0.4466      0.121        164        640:  78%|███████▊  | 627/809 [12:16<03:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.122     0.4467     0.1209        158        640:  78%|███████▊  | 628/809 [12:17<03:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4469     0.1209        226        640:  78%|███████▊  | 630/809 [12:20<03:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121      0.447     0.1208        232        640:  78%|███████▊  | 632/809 [12:22<03:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4471     0.1207        136        640:  78%|███████▊  | 633/809 [12:23<03:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4472     0.1206        200        640:  78%|███████▊  | 635/809 [12:26<03:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4477     0.1205        168        640:  79%|███████▊  | 637/809 [12:28<03:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4476     0.1205        326        640:  79%|███████▉  | 638/809 [12:29<03:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4477     0.1204        334        640:  79%|███████▉  | 640/809 [12:32<03:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4478     0.1203        151        640:  79%|███████▉  | 642/809 [12:34<03:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118     0.4479     0.1203        273        640:  79%|███████▉  | 643/809 [12:35<03:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118      0.448     0.1202        319        640:  80%|███████▉  | 645/809 [12:38<03:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118     0.4481     0.1202        378        640:  80%|███████▉  | 647/809 [12:40<03:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118      0.448     0.1205        208        640:  80%|████████  | 649/809 [12:42<03:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4478     0.1208        471        640:  80%|████████  | 650/809 [12:43<03:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4479     0.1207        410        640:  81%|████████  | 652/809 [12:46<03:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119      0.448     0.1206        237        640:  81%|████████  | 654/809 [12:48<03:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4479     0.1206        346        640:  81%|████████  | 655/809 [12:49<02:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4478     0.1206        156        640:  81%|████████  | 657/809 [12:52<02:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4477     0.1205        237        640:  81%|████████▏ | 659/809 [12:54<02:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4477     0.1204        138        640:  82%|████████▏ | 661/809 [12:56<02:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4476     0.1205        558        640:  82%|████████▏ | 662/809 [12:57<02:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4477     0.1204        396        640:  82%|████████▏ | 664/809 [13:00<02:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4478     0.1203        260        640:  82%|████████▏ | 666/809 [13:02<02:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4479     0.1202        176        640:  82%|████████▏ | 667/809 [13:03<02:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118     0.4483     0.1201        149        640:  83%|████████▎ | 669/809 [13:06<02:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118     0.4483       0.12        193        640:  83%|████████▎ | 671/809 [13:08<02:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4481       0.12        295        640:  83%|████████▎ | 673/809 [13:10<02:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4482       0.12        137        640:  83%|████████▎ | 674/809 [13:12<02:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4478     0.1201        353        640:  84%|████████▎ | 676/809 [13:14<02:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4475     0.1203        214        640:  84%|████████▍ | 678/809 [13:16<02:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4476     0.1203        195        640:  84%|████████▍ | 679/809 [13:18<02:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4477     0.1202        177        640:  84%|████████▍ | 681/809 [13:20<02:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4477     0.1201        239        640:  84%|████████▍ | 683/809 [13:22<02:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4477       0.12        187        640:  85%|████████▍ | 684/809 [13:23<02:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4477     0.1199        120        640:  85%|████████▍ | 686/809 [13:26<02:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4478     0.1198        187        640:  85%|████████▌ | 688/809 [13:28<02:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4478     0.1198        264        640:  85%|████████▌ | 690/809 [13:30<02:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4478     0.1197        346        640:  85%|████████▌ | 691/809 [13:31<02:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118     0.4479     0.1196        312        640:  86%|████████▌ | 693/809 [13:34<02:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117      0.448     0.1195        208        640:  86%|████████▌ | 695/809 [13:36<02:11,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117      0.448     0.1195        199        640:  86%|████████▌ | 697/809 [13:38<02:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117      0.448     0.1194        339        640:  86%|████████▋ | 698/809 [13:40<02:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4481     0.1194        163        640:  87%|████████▋ | 700/809 [13:42<02:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.116     0.4481     0.1194        424        640:  87%|████████▋ | 702/809 [13:44<02:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.116     0.4481     0.1193        183        640:  87%|████████▋ | 703/809 [13:45<02:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.116     0.4481     0.1195         89        640:  87%|████████▋ | 705/809 [13:48<02:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.116      0.448     0.1195        302        640:  87%|████████▋ | 707/809 [13:50<01:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.116     0.4479     0.1195        207        640:  88%|████████▊ | 709/809 [13:53<01:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.116     0.4479     0.1195        264        640:  88%|████████▊ | 710/809 [13:54<01:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4478     0.1197        524        640:  88%|████████▊ | 712/809 [13:56<01:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4477     0.1198        231        640:  88%|████████▊ | 714/809 [13:58<01:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4478     0.1198        191        640:  88%|████████▊ | 715/809 [14:00<01:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4477     0.1197        260        640:  89%|████████▊ | 717/809 [14:02<01:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4477     0.1197        250        640:  89%|████████▉ | 719/809 [14:04<01:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4476     0.1197        345        640:  89%|████████▉ | 720/809 [14:06<01:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4476     0.1197        271        640:  89%|████████▉ | 722/809 [14:08<01:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118     0.4474     0.1198        494        640:  89%|████████▉ | 724/809 [14:10<01:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118     0.4475     0.1197        181        640:  90%|████████▉ | 726/809 [14:13<01:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4476     0.1196        211        640:  90%|████████▉ | 728/809 [14:15<01:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4475     0.1195        347        640:  90%|█████████ | 729/809 [14:16<01:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4474     0.1196        390        640:  90%|█████████ | 731/809 [14:18<01:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118     0.4473     0.1197        413        640:  91%|█████████ | 733/809 [14:21<01:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4474     0.1196        109        640:  91%|█████████ | 734/809 [14:22<01:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4475     0.1196        250        640:  91%|█████████ | 736/809 [14:24<01:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4475     0.1196        272        640:  91%|█████████ | 738/809 [14:27<01:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.116     0.4476     0.1195        430        640:  91%|█████████▏| 740/809 [14:29<01:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.116     0.4476     0.1195        159        640:  92%|█████████▏| 741/809 [14:30<01:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4476     0.1195         61        640:  92%|█████████▏| 743/809 [14:33<01:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4475     0.1194        249        640:  92%|█████████▏| 745/809 [14:35<01:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.117     0.4475     0.1194        333        640:  92%|█████████▏| 747/809 [14:37<01:10,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118     0.4473     0.1197        548        640:  92%|█████████▏| 748/809 [14:38<01:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118     0.4472     0.1197        192        640:  93%|█████████▎| 750/809 [14:41<01:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.118     0.4471     0.1197        302        640:  93%|█████████▎| 752/809 [14:43<01:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119      0.447     0.1197        316        640:  93%|█████████▎| 753/809 [14:44<01:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119      0.447     0.1197        364        640:  93%|█████████▎| 755/809 [14:47<01:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4468     0.1197        264        640:  94%|█████████▎| 757/809 [14:49<01:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4468     0.1197        310        640:  94%|█████████▍| 759/809 [14:51<00:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4468     0.1196        187        640:  94%|█████████▍| 760/809 [14:52<00:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119      0.447     0.1195        211        640:  94%|█████████▍| 762/809 [14:55<00:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.119     0.4467     0.1196        353        640:  94%|█████████▍| 764/809 [14:57<00:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4466     0.1196        355        640:  95%|█████████▍| 765/809 [14:58<00:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4465     0.1196        355        640:  95%|█████████▍| 767/809 [15:01<00:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G       1.12     0.4464     0.1196        287        640:  95%|█████████▌| 769/809 [15:03<00:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4462     0.1195        504        640:  95%|█████████▌| 771/809 [15:05<00:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4461     0.1195        201        640:  95%|█████████▌| 772/809 [15:06<00:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4461     0.1194        258        640:  96%|█████████▌| 774/809 [15:09<00:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4461     0.1194        394        640:  96%|█████████▌| 776/809 [15:11<00:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.122      0.446     0.1194        417        640:  96%|█████████▌| 777/809 [15:12<00:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.122      0.446     0.1194        224        640:  96%|█████████▋| 779/809 [15:14<00:34,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4462     0.1193        116        640:  97%|█████████▋| 781/809 [15:17<00:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.122      0.446     0.1193        301        640:  97%|█████████▋| 783/809 [15:19<00:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.122     0.4459     0.1194        264        640:  97%|█████████▋| 784/809 [15:20<00:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4461     0.1193        219        640:  97%|█████████▋| 786/809 [15:23<00:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4462     0.1192        159        640:  97%|█████████▋| 788/809 [15:25<00:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4461     0.1192        228        640:  98%|█████████▊| 789/809 [15:26<00:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121      0.446     0.1192        102        640:  98%|█████████▊| 791/809 [15:29<00:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4459     0.1191        262        640:  98%|█████████▊| 793/809 [15:31<00:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4459     0.1191        168        640:  98%|█████████▊| 794/809 [15:32<00:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4459      0.119        258        640:  98%|█████████▊| 796/809 [15:35<00:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.121     0.4458      0.119        180        640:  99%|█████████▊| 798/809 [15:37<00:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.122     0.4457      0.119        321        640:  99%|█████████▉| 800/809 [15:39<00:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.122     0.4457      0.119        360        640:  99%|█████████▉| 801/809 [15:40<00:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.122     0.4455      0.119        285        640:  99%|█████████▉| 803/809 [15:43<00:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.122     0.4454      0.119        142        640: 100%|█████████▉| 805/809 [15:45<00:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.122     0.4454      0.119        223        640: 100%|█████████▉| 806/809 [15:46<00:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     81/300      13.6G      1.123     0.4454      0.119        212        640: 100%|█████████▉| 808/809 [15:49<00:01,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   7%|▋         | 5/69 [00:01<00:15,  4.24it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  19%|█▉        | 13/69 [00:03<00:13,  4.08it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  30%|███       | 21/69 [00:05<00:11,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  42%|████▏     | 29/69 [00:07<00:10,  3.84it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  54%|█████▎    | 37/69 [00:09<00:07,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  65%|██████▌   | 45/69 [00:11<00:06,  3.80it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  77%|███████▋  | 53/69 [00:13<00:04,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  88%|████████▊ | 61/69 [00:15<00:01,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward passinit__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  Variable._execution_engine.run_backward(  # Calls into the C++ engine to run the backward passinit__.py:251: UserWarning: grid_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300        14G       1.51     0.3309     0.1709        338        640:   0%|          | 2/809 [00:02<15:32,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4154     0.1283        119        640:   0%|          | 4/809 [00:04<15:35,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.201     0.4103     0.1239        111        640:   1%|          | 6/809 [00:06<15:20,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.162     0.4245     0.1189        253        640:   1%|          | 7/809 [00:08<15:27,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.179     0.4252     0.1234        164        640:   1%|          | 9/809 [00:10<15:18,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.143     0.4337     0.1165        272        640:   1%|▏         | 11/809 [00:12<15:25,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.137     0.4353     0.1151        158        640:   1%|▏         | 12/809 [00:13<15:36,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.154     0.4327     0.1247        518        640:   2%|▏         | 14/809 [00:16<15:30,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4271     0.1393        302        640:   2%|▏         | 16/809 [00:18<15:48,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.185     0.4244     0.1384        281        640:   2%|▏         | 17/809 [00:19<16:01,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4189     0.1389        465        640:   2%|▏         | 19/809 [00:22<16:01,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.192     0.4218     0.1377        275        640:   3%|▎         | 21/809 [00:24<16:18,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.199     0.4187     0.1367        165        640:   3%|▎         | 22/809 [00:26<16:01,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.226     0.4119     0.1373        302        640:   3%|▎         | 24/809 [00:28<15:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.227     0.4115     0.1366        100        640:   3%|▎         | 26/809 [00:30<15:27,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.239     0.4075     0.1343        254        640:   3%|▎         | 28/809 [00:33<15:20,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.232     0.4087     0.1328        257        640:   4%|▎         | 29/809 [00:34<15:13,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.232     0.4097     0.1365        234        640:   4%|▍         | 31/809 [00:36<15:17,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4211     0.1324        133        640:   4%|▍         | 33/809 [00:38<15:02,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4215      0.132        398        640:   4%|▍         | 34/809 [00:40<14:59,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.169     0.4275     0.1292        102        640:   4%|▍         | 36/809 [00:42<15:23,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.179      0.426     0.1296        346        640:   5%|▍         | 38/809 [00:44<15:12,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.189      0.422     0.1299        248        640:   5%|▍         | 39/809 [00:46<15:05,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.21     0.4163     0.1329        416        640:   5%|▌         | 41/809 [00:48<15:01,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.217     0.4134     0.1318        288        640:   5%|▌         | 43/809 [00:50<15:07,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.204     0.4164     0.1296        257        640:   6%|▌         | 45/809 [00:53<15:00,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196     0.4187     0.1282        142        640:   6%|▌         | 46/809 [00:54<14:54,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.202     0.4149     0.1276        266        640:   6%|▌         | 48/809 [00:56<14:44,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4167     0.1268        216        640:   6%|▌         | 50/809 [00:58<14:31,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193      0.416      0.126        229        640:   6%|▋         | 52/809 [01:01<14:46,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.184     0.4178     0.1252        144        640:   7%|▋         | 53/809 [01:02<14:41,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.178     0.4193     0.1244        219        640:   7%|▋         | 55/809 [01:04<14:33,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.17      0.421     0.1231        283        640:   7%|▋         | 57/809 [01:06<14:29,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4246     0.1226        104        640:   7%|▋         | 58/809 [01:08<14:45,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4245     0.1237        272        640:   7%|▋         | 60/809 [01:10<14:39,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.159     0.4261     0.1224        220        640:   8%|▊         | 62/809 [01:13<14:56,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.158     0.4267     0.1221        235        640:   8%|▊         | 63/809 [01:14<14:38,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.154     0.4281     0.1218        237        640:   8%|▊         | 65/809 [01:16<14:23,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.157      0.427     0.1223        306        640:   8%|▊         | 67/809 [01:18<14:13,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.151     0.4276     0.1213        151        640:   9%|▊         | 69/809 [01:21<14:22,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.149     0.4284     0.1208        261        640:   9%|▊         | 70/809 [01:22<14:24,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.153     0.4264     0.1207        241        640:   9%|▉         | 72/809 [01:24<14:34,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.153     0.4259     0.1201        272        640:   9%|▉         | 74/809 [01:26<14:21,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.15     0.4268     0.1199        205        640:   9%|▉         | 75/809 [01:28<14:16,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.145     0.4282     0.1186        258        640:  10%|▉         | 77/809 [01:30<14:19,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.144     0.4289     0.1188        168        640:  10%|▉         | 79/809 [01:32<14:11,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.144     0.4288     0.1186        280        640:  10%|█         | 81/809 [01:35<14:22,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.15     0.4275     0.1221        476        640:  10%|█         | 82/809 [01:36<14:29,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.154      0.426      0.124        542        640:  10%|█         | 84/809 [01:38<14:41,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.146     0.4277      0.123        219        640:  11%|█         | 86/809 [01:41<14:14,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.143     0.4286      0.122        222        640:  11%|█         | 88/809 [01:43<14:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.142     0.4292     0.1221        445        640:  11%|█         | 90/809 [01:45<14:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.138     0.4302     0.1217        199        640:  11%|█         | 91/809 [01:47<14:01,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.144     0.4278     0.1225        388        640:  11%|█▏        | 93/809 [01:49<14:05,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.151     0.4259      0.124        366        640:  12%|█▏        | 95/809 [01:51<14:00,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.151     0.4265     0.1242        260        640:  12%|█▏        | 96/809 [01:52<13:58,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.149     0.4267     0.1233        239        640:  12%|█▏        | 98/809 [01:55<14:03,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.15     0.4265     0.1229        215        640:  12%|█▏        | 100/809 [01:57<14:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.148     0.4272     0.1225        255        640:  12%|█▏        | 101/809 [01:58<14:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.147     0.4287     0.1221        230        640:  13%|█▎        | 103/809 [02:01<13:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.145     0.4291     0.1215        125        640:  13%|█▎        | 105/809 [02:03<13:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.139     0.4318     0.1207         57        640:  13%|█▎        | 107/809 [02:05<13:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.138     0.4327     0.1207        271        640:  13%|█▎        | 108/809 [02:07<13:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.139     0.4327     0.1202        277        640:  14%|█▎        | 110/809 [02:09<13:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.14     0.4322     0.1199        346        640:  14%|█▍        | 112/809 [02:11<13:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.141     0.4321     0.1197        275        640:  14%|█▍        | 113/809 [02:13<13:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.146     0.4303     0.1197        240        640:  14%|█▍        | 115/809 [02:15<13:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.149     0.4293     0.1201        242        640:  14%|█▍        | 117/809 [02:17<13:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.147     0.4299     0.1197        143        640:  15%|█▍        | 118/809 [02:18<13:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.146     0.4305     0.1206        215        640:  15%|█▍        | 120/809 [02:21<13:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.144     0.4314     0.1201        169        640:  15%|█▌        | 122/809 [02:23<13:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.14     0.4321     0.1198        283        640:  15%|█▌        | 123/809 [02:24<13:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.136     0.4329     0.1193         85        640:  15%|█▌        | 125/809 [02:27<13:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.135     0.4336     0.1192        400        640:  16%|█▌        | 127/809 [02:29<13:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.134     0.4348      0.119        298        640:  16%|█▌        | 129/809 [02:31<13:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.135     0.4348     0.1194        275        640:  16%|█▌        | 130/809 [02:33<13:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.141     0.4335     0.1213        588        640:  16%|█▋        | 132/809 [02:35<13:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.14     0.4339     0.1213        258        640:  16%|█▋        | 133/809 [02:36<13:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.14     0.4346     0.1213        193        640:  17%|█▋        | 135/809 [02:39<13:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.139     0.4347     0.1211        137        640:  17%|█▋        | 137/809 [02:41<13:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.141     0.4344     0.1216        331        640:  17%|█▋        | 138/809 [02:42<13:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.144      0.434     0.1215        240        640:  17%|█▋        | 140/809 [02:45<13:41,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.144     0.4339     0.1214        314        640:  18%|█▊        | 142/809 [02:47<13:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.143     0.4341     0.1213        234        640:  18%|█▊        | 143/809 [02:48<13:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.141     0.4344      0.121        213        640:  18%|█▊        | 145/809 [02:51<13:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.138     0.4352     0.1205        101        640:  18%|█▊        | 147/809 [02:53<13:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.139     0.4349     0.1204        269        640:  18%|█▊        | 149/809 [02:56<13:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.141     0.4343     0.1209        554        640:  19%|█▊        | 150/809 [02:57<13:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.141     0.4342     0.1207        333        640:  19%|█▉        | 152/809 [02:59<13:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.139     0.4352     0.1203        253        640:  19%|█▉        | 154/809 [03:01<12:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.138     0.4354       0.12        173        640:  19%|█▉        | 155/809 [03:03<12:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.139     0.4347     0.1198        389        640:  19%|█▉        | 157/809 [03:05<12:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.14     0.4344     0.1197        389        640:  20%|█▉        | 159/809 [03:07<12:28,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.14     0.4343     0.1195        253        640:  20%|█▉        | 161/809 [03:10<12:20,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.141     0.4342     0.1199        328        640:  20%|██        | 162/809 [03:11<12:19,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.144     0.4331     0.1203        409        640:  20%|██        | 164/809 [03:13<12:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.144     0.4335     0.1202        273        640:  21%|██        | 166/809 [03:15<12:18,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.144     0.4335     0.1199        352        640:  21%|██        | 168/809 [03:18<12:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.143     0.4342     0.1197        288        640:  21%|██        | 169/809 [03:19<12:17,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.143     0.4344     0.1197        253        640:  21%|██        | 171/809 [03:21<12:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.141      0.435     0.1193        126        640:  21%|██▏       | 173/809 [03:23<12:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.14     0.4352     0.1193        162        640:  22%|██▏       | 174/809 [03:25<12:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.138     0.4358     0.1192        121        640:  22%|██▏       | 176/809 [03:27<12:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.138     0.4361     0.1192        284        640:  22%|██▏       | 178/809 [03:29<12:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.138     0.4361     0.1191        370        640:  22%|██▏       | 180/809 [03:32<12:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.136     0.4371     0.1188        184        640:  22%|██▏       | 181/809 [03:33<12:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.134     0.4382     0.1185        145        640:  23%|██▎       | 183/809 [03:35<12:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.133     0.4388     0.1185        302        640:  23%|██▎       | 185/809 [03:37<12:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.133     0.4392     0.1185        386        640:  23%|██▎       | 187/809 [03:40<11:57,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.133     0.4392     0.1186        344        640:  23%|██▎       | 188/809 [03:41<12:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.134     0.4391     0.1186        521        640:  23%|██▎       | 190/809 [03:43<12:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.136     0.4386     0.1189        516        640:  24%|██▎       | 192/809 [03:46<12:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.135     0.4393     0.1187        253        640:  24%|██▍       | 193/809 [03:47<12:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.135     0.4393     0.1186        259        640:  24%|██▍       | 195/809 [03:49<12:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.135     0.4398     0.1184        305        640:  24%|██▍       | 197/809 [03:52<12:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.136     0.4394     0.1187        518        640:  24%|██▍       | 198/809 [03:53<12:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.136     0.4396     0.1193        237        640:  25%|██▍       | 200/809 [03:55<12:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.134     0.4403     0.1189        282        640:  25%|██▍       | 202/809 [03:58<11:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.134     0.4405     0.1194        229        640:  25%|██▌       | 203/809 [03:59<11:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.133     0.4411     0.1192        200        640:  25%|██▌       | 205/809 [04:01<11:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.133     0.4414      0.119        256        640:  26%|██▌       | 207/809 [04:03<11:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.132     0.4416     0.1191        243        640:  26%|██▌       | 208/809 [04:05<11:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.137     0.4405     0.1216        491        640:  26%|██▌       | 210/809 [04:07<11:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.139     0.4399     0.1217        287        640:  26%|██▌       | 212/809 [04:09<11:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.14     0.4402     0.1216        318        640:  26%|██▋       | 213/809 [04:11<11:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.141     0.4399     0.1235        211        640:  27%|██▋       | 215/809 [04:13<11:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.144      0.439     0.1243        746        640:  27%|██▋       | 217/809 [04:15<11:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.143     0.4394      0.124        111        640:  27%|██▋       | 219/809 [04:18<11:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.143     0.4394      0.124        236        640:  27%|██▋       | 220/809 [04:19<11:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.146     0.4386     0.1247        330        640:  27%|██▋       | 222/809 [04:21<11:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.148     0.4382      0.125        233        640:  28%|██▊       | 224/809 [04:24<11:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.147     0.4382     0.1249        159        640:  28%|██▊       | 226/809 [04:26<11:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.149     0.4377     0.1254        426        640:  28%|██▊       | 227/809 [04:27<11:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.149     0.4375     0.1254        298        640:  28%|██▊       | 229/809 [04:29<11:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.149     0.4377     0.1253        156        640:  29%|██▊       | 231/809 [04:32<11:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.149     0.4379     0.1252         81        640:  29%|██▊       | 232/809 [04:33<11:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.151     0.4374     0.1256        328        640:  29%|██▉       | 234/809 [04:35<11:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.151     0.4374     0.1255        167        640:  29%|██▉       | 236/809 [04:38<11:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.151     0.4376     0.1254        146        640:  29%|██▉       | 237/809 [04:39<11:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.151     0.4374     0.1255        143        640:  30%|██▉       | 239/809 [04:41<11:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.153     0.4364     0.1259        416        640:  30%|██▉       | 241/809 [04:44<11:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.154      0.436     0.1259        160        640:  30%|██▉       | 242/809 [04:45<11:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.154     0.4358     0.1259        151        640:  30%|███       | 244/809 [04:47<11:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.155     0.4354     0.1263        260        640:  30%|███       | 246/809 [04:49<10:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.156     0.4352     0.1265        336        640:  31%|███       | 248/809 [04:52<10:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.156     0.4355     0.1263        197        640:  31%|███       | 250/809 [04:54<10:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.156     0.4353     0.1264        249        640:  31%|███       | 251/809 [04:55<10:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.155     0.4357     0.1262        222        640:  31%|███▏      | 253/809 [04:58<10:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.156     0.4354     0.1261        220        640:  32%|███▏      | 255/809 [05:00<10:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.155     0.4355     0.1259        150        640:  32%|███▏      | 257/809 [05:02<10:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.155     0.4356     0.1259        285        640:  32%|███▏      | 258/809 [05:03<10:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.159     0.4344     0.1277        723        640:  32%|███▏      | 260/809 [05:06<11:25,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.162     0.4335     0.1281        278        640:  32%|███▏      | 262/809 [05:08<10:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.163     0.4333     0.1281        219        640:  33%|███▎      | 263/809 [05:10<10:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166     0.4326     0.1285        247        640:  33%|███▎      | 265/809 [05:12<10:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.168     0.4324     0.1286        109        640:  33%|███▎      | 267/809 [05:14<10:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166     0.4325     0.1285        130        640:  33%|███▎      | 268/809 [05:15<10:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.167     0.4322     0.1288        250        640:  33%|███▎      | 270/809 [05:18<10:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.17     0.4315     0.1297        574        640:  34%|███▎      | 272/809 [05:20<10:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.169     0.4314     0.1296        206        640:  34%|███▎      | 273/809 [05:21<10:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.17     0.4312     0.1299        222        640:  34%|███▍      | 275/809 [05:24<10:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.17     0.4312     0.1297        202        640:  34%|███▍      | 277/809 [05:26<10:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.172     0.4306     0.1303        214        640:  34%|███▍      | 279/809 [05:28<10:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.171     0.4309     0.1303         60        640:  35%|███▍      | 280/809 [05:30<10:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.175     0.4301     0.1313        388        640:  35%|███▍      | 282/809 [05:32<10:04,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.175     0.4301     0.1314        112        640:  35%|███▌      | 284/809 [05:34<10:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.177     0.4294     0.1314        312        640:  35%|███▌      | 286/809 [05:37<10:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.177     0.4294     0.1313        238        640:  35%|███▌      | 287/809 [05:38<10:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18     0.4286     0.1324        226        640:  36%|███▌      | 289/809 [05:40<10:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.182      0.428     0.1324        273        640:  36%|███▌      | 291/809 [05:42<10:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.182      0.428     0.1323         92        640:  36%|███▌      | 292/809 [05:44<10:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.183     0.4275     0.1322        380        640:  36%|███▋      | 294/809 [05:46<10:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.184     0.4272     0.1324        280        640:  37%|███▋      | 296/809 [05:48<10:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.183     0.4273     0.1322        124        640:  37%|███▋      | 297/809 [05:50<10:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.185     0.4269     0.1322        262        640:  37%|███▋      | 299/809 [05:52<10:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.186     0.4264     0.1322        324        640:  37%|███▋      | 301/809 [05:54<09:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.187     0.4262     0.1324        241        640:  37%|███▋      | 302/809 [05:55<09:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.189     0.4258     0.1323        217        640:  38%|███▊      | 304/809 [05:58<09:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4254     0.1328        587        640:  38%|███▊      | 306/809 [06:00<10:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.192     0.4253     0.1328        143        640:  38%|███▊      | 307/809 [06:01<10:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4247      0.133        504        640:  38%|███▊      | 309/809 [06:04<09:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4242     0.1329        256        640:  38%|███▊      | 311/809 [06:06<09:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196     0.4242     0.1329        337        640:  39%|███▊      | 312/809 [06:07<09:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.197     0.4239      0.133        332        640:  39%|███▉      | 314/809 [06:10<09:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4235     0.1329        157        640:  39%|███▉      | 316/809 [06:12<09:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4237     0.1328        194        640:  39%|███▉      | 317/809 [06:14<09:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4235     0.1326        393        640:  39%|███▉      | 319/809 [06:16<09:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4236     0.1325        182        640:  40%|███▉      | 321/809 [06:18<09:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.199     0.4232     0.1326        250        640:  40%|███▉      | 323/809 [06:20<09:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4235     0.1324        108        640:  40%|████      | 324/809 [06:22<09:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4234     0.1323        251        640:  40%|████      | 326/809 [06:24<09:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4236     0.1321        281        640:  41%|████      | 328/809 [06:26<09:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.199     0.4233     0.1322        395        640:  41%|████      | 329/809 [06:28<09:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.197     0.4239     0.1319        132        640:  41%|████      | 331/809 [06:30<09:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196     0.4243     0.1317        287        640:  41%|████      | 333/809 [06:32<09:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4248     0.1317        214        640:  41%|████▏     | 335/809 [06:35<09:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4248     0.1315        227        640:  42%|████▏     | 336/809 [06:36<09:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4254     0.1313        381        640:  42%|████▏     | 338/809 [06:38<09:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4253     0.1316        158        640:  42%|████▏     | 340/809 [06:41<09:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4253     0.1316        282        640:  42%|████▏     | 341/809 [06:42<09:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4252     0.1315        212        640:  42%|████▏     | 343/809 [06:44<09:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4248     0.1317        350        640:  43%|████▎     | 345/809 [06:46<09:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195      0.425     0.1315        179        640:  43%|████▎     | 347/809 [06:49<08:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4252     0.1315        191        640:  43%|████▎     | 348/809 [06:50<09:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4251     0.1313        196        640:  43%|████▎     | 350/809 [06:52<08:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4253     0.1312        102        640:  44%|████▎     | 352/809 [06:55<08:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4251     0.1313        303        640:  44%|████▎     | 353/809 [06:56<08:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4254     0.1311        210        640:  44%|████▍     | 355/809 [06:58<08:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4248     0.1314        399        640:  44%|████▍     | 357/809 [07:00<08:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4248     0.1314        177        640:  44%|████▍     | 358/809 [07:02<08:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4252     0.1312        106        640:  44%|████▍     | 360/809 [07:04<08:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4252     0.1313        378        640:  45%|████▍     | 362/809 [07:06<08:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196      0.425     0.1313        213        640:  45%|████▍     | 364/809 [07:09<08:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196      0.425     0.1313        244        640:  45%|████▌     | 365/809 [07:10<08:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196     0.4249     0.1313        108        640:  45%|████▌     | 367/809 [07:12<08:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4244     0.1318        333        640:  46%|████▌     | 369/809 [07:15<08:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4245     0.1318        217        640:  46%|████▌     | 370/809 [07:16<08:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.199     0.4242      0.132        135        640:  46%|████▌     | 372/809 [07:18<08:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.199     0.4242      0.132        368        640:  46%|████▌     | 374/809 [07:21<08:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.199     0.4243     0.1319        225        640:  46%|████▋     | 375/809 [07:22<08:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G        1.2     0.4238     0.1323        290        640:  47%|████▋     | 377/809 [07:24<08:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.201     0.4237     0.1324        293        640:  47%|████▋     | 379/809 [07:26<08:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.201     0.4237     0.1324        229        640:  47%|████▋     | 381/809 [07:29<08:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G        1.2     0.4237     0.1323        141        640:  47%|████▋     | 382/809 [07:30<08:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.201     0.4235     0.1325        398        640:  47%|████▋     | 384/809 [07:32<08:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.201     0.4234     0.1325        200        640:  48%|████▊     | 386/809 [07:35<08:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.202     0.4232     0.1325        306        640:  48%|████▊     | 387/809 [07:36<08:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.201     0.4233     0.1325        238        640:  48%|████▊     | 389/809 [07:38<08:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.201     0.4235     0.1325        225        640:  48%|████▊     | 391/809 [07:41<08:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G        1.2     0.4237     0.1325        216        640:  49%|████▊     | 393/809 [07:43<08:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G        1.2     0.4239     0.1324        229        640:  49%|████▊     | 394/809 [07:44<08:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.199     0.4242     0.1322        312        640:  49%|████▉     | 396/809 [07:46<08:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.198     0.4248      0.132        212        640:  49%|████▉     | 398/809 [07:49<08:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196     0.4254     0.1317        152        640:  49%|████▉     | 400/809 [07:51<07:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4259     0.1318        278        640:  50%|████▉     | 402/809 [07:53<07:47,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4263     0.1317        109        640:  50%|████▉     | 403/809 [07:54<07:44,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4268     0.1315        144        640:  50%|█████     | 405/809 [07:57<07:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4274     0.1314         63        640:  50%|█████     | 407/809 [07:59<07:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4275     0.1314        309        640:  50%|█████     | 408/809 [08:00<07:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.19     0.4281     0.1312        155        640:  51%|█████     | 410/809 [08:03<07:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4278     0.1316        342        640:  51%|█████     | 412/809 [08:05<07:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191      0.428     0.1314        259        640:  51%|█████     | 414/809 [08:07<07:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191      0.428     0.1314        250        640:  51%|█████▏    | 415/809 [08:09<08:00,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191      0.428     0.1313        190        640:  52%|█████▏    | 417/809 [08:11<07:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.19     0.4281     0.1311        270        640:  52%|█████▏    | 419/809 [08:13<07:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.19      0.428     0.1311        224        640:  52%|█████▏    | 420/809 [08:15<07:57,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.19     0.4281     0.1311        356        640:  52%|█████▏    | 422/809 [08:17<07:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.19     0.4282     0.1311        354        640:  52%|█████▏    | 424/809 [08:19<07:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4281     0.1311        360        640:  53%|█████▎    | 425/809 [08:21<07:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.19     0.4286      0.131        295        640:  53%|█████▎    | 427/809 [08:23<07:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.19     0.4287     0.1309        185        640:  53%|█████▎    | 429/809 [08:25<07:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4283     0.1311        316        640:  53%|█████▎    | 430/809 [08:26<07:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4276     0.1321        589        640:  53%|█████▎    | 432/809 [08:29<07:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4276     0.1321        231        640:  54%|█████▎    | 434/809 [08:31<07:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196     0.4271     0.1322        553        640:  54%|█████▍    | 436/809 [08:33<07:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4272     0.1321        105        640:  54%|█████▍    | 437/809 [08:35<07:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4275      0.132        266        640:  54%|█████▍    | 439/809 [08:37<07:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4277     0.1318        278        640:  55%|█████▍    | 441/809 [08:39<07:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4278      0.132        117        640:  55%|█████▍    | 443/809 [08:42<07:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4277      0.132        407        640:  55%|█████▍    | 444/809 [08:43<07:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4279     0.1318        266        640:  55%|█████▌    | 446/809 [08:45<07:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4282     0.1316        149        640:  55%|█████▌    | 448/809 [08:47<07:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4284     0.1316        190        640:  56%|█████▌    | 449/809 [08:49<06:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.192     0.4288     0.1314        186        640:  56%|█████▌    | 451/809 [08:51<06:53,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4293     0.1312         90        640:  56%|█████▌    | 453/809 [08:53<06:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.19     0.4296     0.1312        339        640:  56%|█████▌    | 455/809 [08:55<06:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4294     0.1315        270        640:  56%|█████▋    | 456/809 [08:57<06:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4296     0.1314        163        640:  57%|█████▋    | 458/809 [08:59<06:43,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4298     0.1313        286        640:  57%|█████▋    | 460/809 [09:01<06:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4298     0.1312        439        640:  57%|█████▋    | 462/809 [09:04<06:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191       0.43     0.1311        175        640:  57%|█████▋    | 463/809 [09:05<06:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191       0.43     0.1311        175        640:  57%|█████▋    | 465/809 [09:07<06:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191       0.43     0.1311        151        640:  58%|█████▊    | 467/809 [09:09<06:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.19     0.4301      0.131        236        640:  58%|█████▊    | 468/809 [09:11<06:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4299      0.131        207        640:  58%|█████▊    | 470/809 [09:13<06:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4294     0.1313        528        640:  58%|█████▊    | 472/809 [09:15<06:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4293     0.1312        370        640:  59%|█████▊    | 474/809 [09:18<06:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4289     0.1314        283        640:  59%|█████▊    | 475/809 [09:19<06:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4287     0.1314        149        640:  59%|█████▉    | 477/809 [09:21<06:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196     0.4288     0.1316        328        640:  59%|█████▉    | 479/809 [09:24<06:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196     0.4289     0.1316        196        640:  59%|█████▉    | 480/809 [09:25<06:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4293     0.1314        230        640:  60%|█████▉    | 482/809 [09:27<06:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4294     0.1313        112        640:  60%|█████▉    | 484/809 [09:29<06:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4298     0.1311        117        640:  60%|██████    | 486/809 [09:32<06:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4299     0.1311        256        640:  60%|██████    | 487/809 [09:33<06:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4298     0.1311        272        640:  60%|██████    | 489/809 [09:35<06:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4298      0.131        494        640:  61%|██████    | 491/809 [09:37<06:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4296     0.1311        261        640:  61%|██████    | 492/809 [09:39<06:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4296     0.1311        178        640:  61%|██████    | 494/809 [09:41<06:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4294      0.131        284        640:  61%|██████▏   | 496/809 [09:43<06:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4294     0.1309        232        640:  62%|██████▏   | 498/809 [09:46<06:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4293     0.1309        350        640:  62%|██████▏   | 499/809 [09:47<06:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4293     0.1309        179        640:  62%|██████▏   | 501/809 [09:49<06:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196      0.429      0.131        265        640:  62%|██████▏   | 503/809 [09:52<06:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196     0.4291      0.131        201        640:  62%|██████▏   | 504/809 [09:53<05:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196     0.4291      0.131        272        640:  63%|██████▎   | 506/809 [09:55<05:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4293     0.1309        200        640:  63%|██████▎   | 508/809 [09:58<05:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4295     0.1308        134        640:  63%|██████▎   | 509/809 [09:59<05:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4293     0.1307        322        640:  63%|██████▎   | 511/809 [10:01<05:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.196     0.4291     0.1308        452        640:  63%|██████▎   | 513/809 [10:03<05:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4295     0.1306        133        640:  64%|██████▎   | 515/809 [10:06<05:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4296     0.1306        226        640:  64%|██████▍   | 516/809 [10:07<05:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194     0.4294     0.1305        313        640:  64%|██████▍   | 518/809 [10:09<05:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4292     0.1306        399        640:  64%|██████▍   | 520/809 [10:12<05:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.195     0.4292     0.1305        193        640:  64%|██████▍   | 521/809 [10:13<05:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4296     0.1303        188        640:  65%|██████▍   | 523/809 [10:15<05:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.192     0.4298     0.1301        153        640:  65%|██████▍   | 525/809 [10:17<05:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4296     0.1307        539        640:  65%|██████▌   | 526/809 [10:19<05:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4295     0.1306        147        640:  65%|██████▌   | 528/809 [10:21<05:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.192     0.4298     0.1304        448        640:  66%|██████▌   | 530/809 [10:23<05:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191       0.43     0.1303        245        640:  66%|██████▌   | 532/809 [10:26<05:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191     0.4301     0.1302        131        640:  66%|██████▌   | 533/809 [10:27<05:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.192     0.4297     0.1303        212        640:  66%|██████▌   | 535/809 [10:29<05:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.192     0.4298     0.1302        176        640:  66%|██████▋   | 537/809 [10:32<05:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.192     0.4296     0.1302        293        640:  67%|██████▋   | 538/809 [10:33<05:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.194      0.429     0.1305        395        640:  67%|██████▋   | 540/809 [10:35<05:25,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4293     0.1303        274        640:  67%|██████▋   | 542/809 [10:38<05:27,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4292     0.1303        295        640:  67%|██████▋   | 544/809 [10:40<05:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193      0.429     0.1305        363        640:  67%|██████▋   | 546/809 [10:42<05:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4291     0.1304        227        640:  68%|██████▊   | 547/809 [10:44<05:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4292     0.1303        295        640:  68%|██████▊   | 549/809 [10:46<05:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4293     0.1303        240        640:  68%|██████▊   | 551/809 [10:48<05:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4292     0.1303        375        640:  68%|██████▊   | 552/809 [10:49<05:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.193     0.4295     0.1302        216        640:  68%|██████▊   | 554/809 [10:52<04:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.192     0.4297     0.1301        118        640:  69%|██████▊   | 556/809 [10:54<05:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.192     0.4298     0.1301        135        640:  69%|██████▉   | 557/809 [10:55<05:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.191       0.43     0.1301        303        640:  69%|██████▉   | 559/809 [10:58<04:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.189     0.4303     0.1299        202        640:  69%|██████▉   | 561/809 [11:00<04:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.188     0.4308     0.1297         67        640:  70%|██████▉   | 563/809 [11:02<04:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.187     0.4309     0.1297        240        640:  70%|██████▉   | 564/809 [11:04<04:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.188     0.4308     0.1297        315        640:  70%|██████▉   | 566/809 [11:06<04:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.188      0.431     0.1297        251        640:  70%|███████   | 568/809 [11:08<04:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.187     0.4311     0.1296        272        640:  70%|███████   | 570/809 [11:10<04:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.187      0.431     0.1296        223        640:  71%|███████   | 571/809 [11:12<04:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.187     0.4309     0.1296        245        640:  71%|███████   | 573/809 [11:14<04:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.186     0.4311     0.1294        238        640:  71%|███████   | 575/809 [11:16<04:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.186     0.4311     0.1295        267        640:  71%|███████   | 576/809 [11:17<04:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.186     0.4311     0.1295        297        640:  71%|███████▏  | 578/809 [11:20<04:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.186     0.4312     0.1294        256        640:  72%|███████▏  | 580/809 [11:22<04:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.186      0.431     0.1295        229        640:  72%|███████▏  | 582/809 [11:25<04:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.187     0.4308     0.1295        235        640:  72%|███████▏  | 583/809 [11:26<04:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.187     0.4308     0.1295         96        640:  72%|███████▏  | 585/809 [11:28<04:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.187     0.4309     0.1294        304        640:  73%|███████▎  | 587/809 [11:30<04:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.187     0.4307     0.1295        254        640:  73%|███████▎  | 588/809 [11:32<04:31,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.187     0.4309     0.1294        161        640:  73%|███████▎  | 590/809 [11:34<04:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.186     0.4312     0.1293        258        640:  73%|███████▎  | 592/809 [11:36<04:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.186     0.4313     0.1292         82        640:  73%|███████▎  | 593/809 [11:38<04:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.184     0.4315     0.1291         80        640:  74%|███████▎  | 595/809 [11:40<04:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.184     0.4317      0.129        159        640:  74%|███████▍  | 597/809 [11:42<04:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.183     0.4317     0.1289        318        640:  74%|███████▍  | 598/809 [11:44<04:16,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.183     0.4319     0.1288        265        640:  74%|███████▍  | 600/809 [11:46<04:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.182      0.432     0.1286        201        640:  74%|███████▍  | 602/809 [11:48<04:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.182     0.4321     0.1285        149        640:  75%|███████▍  | 603/809 [11:49<04:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4323     0.1284        128        640:  75%|███████▍  | 605/809 [11:52<04:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18     0.4325     0.1283        110        640:  75%|███████▌  | 607/809 [11:54<04:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18     0.4325     0.1284        395        640:  75%|███████▌  | 608/809 [11:56<04:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18     0.4325     0.1283        401        640:  75%|███████▌  | 610/809 [11:58<03:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.179     0.4326     0.1281        161        640:  76%|███████▌  | 612/809 [12:00<04:02,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18     0.4324     0.1282        325        640:  76%|███████▌  | 613/809 [12:02<04:00,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4321     0.1283        484        640:  76%|███████▌  | 615/809 [12:04<04:00,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4323     0.1282        109        640:  76%|███████▋  | 617/809 [12:06<03:51,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4323     0.1282        231        640:  76%|███████▋  | 618/809 [12:08<03:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4323     0.1282        299        640:  77%|███████▋  | 620/809 [12:10<03:53,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4322     0.1281        112        640:  77%|███████▋  | 622/809 [12:13<03:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4321     0.1281        203        640:  77%|███████▋  | 623/809 [12:14<03:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18     0.4321      0.128        261        640:  77%|███████▋  | 625/809 [12:16<03:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18      0.432     0.1279        171        640:  78%|███████▊  | 627/809 [12:18<03:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18     0.4321     0.1279        176        640:  78%|███████▊  | 628/809 [12:20<03:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18      0.432     0.1278        188        640:  78%|███████▊  | 630/809 [12:22<03:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18     0.4319     0.1278        261        640:  78%|███████▊  | 632/809 [12:24<03:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4318     0.1279        283        640:  78%|███████▊  | 634/809 [12:27<03:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4319     0.1278         80        640:  78%|███████▊  | 635/809 [12:28<03:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4318     0.1278        364        640:  79%|███████▊  | 637/809 [12:30<03:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4318     0.1276        174        640:  79%|███████▉  | 639/809 [12:33<03:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18      0.432     0.1276        174        640:  79%|███████▉  | 640/809 [12:34<03:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18     0.4319     0.1275        288        640:  79%|███████▉  | 642/809 [12:36<03:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18     0.4318     0.1275        296        640:  80%|███████▉  | 644/809 [12:38<03:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4317     0.1275        228        640:  80%|███████▉  | 645/809 [12:40<03:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4315     0.1275        275        640:  80%|███████▉  | 647/809 [12:42<03:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.182     0.4312     0.1275        442        640:  80%|████████  | 649/809 [12:44<03:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4315     0.1274        143        640:  80%|████████  | 651/809 [12:47<03:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4316     0.1275        177        640:  81%|████████  | 652/809 [12:48<03:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.181     0.4318     0.1273        274        640:  81%|████████  | 654/809 [12:50<03:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18      0.432     0.1272        315        640:  81%|████████  | 656/809 [12:53<03:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.18     0.4322     0.1272        166        640:  81%|████████  | 657/809 [12:54<02:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.179     0.4323     0.1271        253        640:  81%|████████▏ | 659/809 [12:56<02:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.178     0.4326      0.127        110        640:  82%|████████▏ | 661/809 [12:59<02:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.177     0.4329     0.1269         96        640:  82%|████████▏ | 662/809 [13:01<02:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.177     0.4328     0.1269        207        640:  82%|████████▏ | 664/809 [13:02<02:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.176     0.4329     0.1268        205        640:  82%|████████▏ | 666/809 [13:04<02:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.176     0.4328     0.1267        178        640:  83%|████████▎ | 668/809 [13:07<02:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.175     0.4331     0.1267        127        640:  83%|████████▎ | 669/809 [13:08<02:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.174     0.4335     0.1266        287        640:  83%|████████▎ | 671/809 [13:10<02:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.174     0.4339     0.1265        119        640:  83%|████████▎ | 673/809 [13:12<02:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.172     0.4343     0.1263        208        640:  83%|████████▎ | 675/809 [13:15<02:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.172     0.4344     0.1262        189        640:  84%|████████▎ | 676/809 [13:16<02:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.171     0.4344     0.1263        221        640:  84%|████████▍ | 678/809 [13:18<02:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.171     0.4346     0.1262        241        640:  84%|████████▍ | 680/809 [13:21<02:28,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.171     0.4346     0.1261        237        640:  84%|████████▍ | 682/809 [13:23<02:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.171     0.4345     0.1261        493        640:  84%|████████▍ | 683/809 [13:24<02:25,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.171     0.4347      0.126        347        640:  85%|████████▍ | 685/809 [13:26<02:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.17     0.4348      0.126        238        640:  85%|████████▍ | 687/809 [13:29<02:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.17      0.435      0.126        205        640:  85%|████████▌ | 689/809 [13:31<02:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.169     0.4352     0.1259        100        640:  85%|████████▌ | 691/809 [13:33<02:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.169     0.4352     0.1258        216        640:  86%|████████▌ | 692/809 [13:35<02:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.169     0.4353     0.1258        241        640:  86%|████████▌ | 694/809 [13:37<02:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.169     0.4352     0.1259        312        640:  86%|████████▌ | 696/809 [13:39<02:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.169     0.4351     0.1259        270        640:  86%|████████▌ | 697/809 [13:41<02:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.17     0.4349     0.1259        173        640:  86%|████████▋ | 699/809 [13:43<02:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.17     0.4349     0.1259        264        640:  87%|████████▋ | 701/809 [13:45<02:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G       1.17      0.435      0.126        309        640:  87%|████████▋ | 702/809 [13:46<02:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.169     0.4351      0.126        368        640:  87%|████████▋ | 704/809 [13:49<02:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.168     0.4353     0.1259        169        640:  87%|████████▋ | 706/809 [13:51<02:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.168     0.4354     0.1259        410        640:  88%|████████▊ | 708/809 [13:53<01:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.168     0.4354      0.126        239        640:  88%|████████▊ | 709/809 [13:55<01:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.167     0.4358     0.1259        149        640:  88%|████████▊ | 711/809 [13:57<01:52,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166      0.436     0.1259        252        640:  88%|████████▊ | 713/809 [13:59<01:50,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.165     0.4362     0.1259        193        640:  88%|████████▊ | 715/809 [14:01<01:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.165     0.4364     0.1258        201        640:  89%|████████▊ | 716/809 [14:03<01:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4366     0.1257        233        640:  89%|████████▉ | 718/809 [14:05<01:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4366     0.1257        269        640:  89%|████████▉ | 719/809 [14:06<01:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.165     0.4364     0.1265        421        640:  89%|████████▉ | 721/809 [14:09<01:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.165     0.4365     0.1265        272        640:  89%|████████▉ | 723/809 [14:11<01:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.165     0.4364     0.1264        161        640:  90%|████████▉ | 725/809 [14:13<01:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.165     0.4362     0.1267        656        640:  90%|████████▉ | 726/809 [14:15<01:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.165     0.4365     0.1266        191        640:  90%|████████▉ | 728/809 [14:17<01:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4366     0.1266        335        640:  90%|█████████ | 730/809 [14:19<01:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.165     0.4363      0.127        499        640:  90%|█████████ | 731/809 [14:20<01:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.165     0.4363      0.127        239        640:  91%|█████████ | 733/809 [14:23<01:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166      0.436     0.1274        340        640:  91%|█████████ | 735/809 [14:25<01:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166     0.4361     0.1273        183        640:  91%|█████████ | 736/809 [14:27<01:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166     0.4362     0.1273        273        640:  91%|█████████ | 738/809 [14:29<01:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166     0.4361     0.1272        359        640:  91%|█████████▏| 740/809 [14:31<01:24,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166      0.436     0.1272        166        640:  92%|█████████▏| 741/809 [14:33<01:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166     0.4361     0.1271        221        640:  92%|█████████▏| 743/809 [14:35<01:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166      0.436     0.1271        344        640:  92%|█████████▏| 745/809 [14:37<01:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.167     0.4359     0.1275        518        640:  92%|█████████▏| 746/809 [14:39<01:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166      0.436     0.1274        212        640:  92%|█████████▏| 748/809 [14:41<01:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166     0.4359     0.1275        318        640:  93%|█████████▎| 750/809 [14:43<01:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166      0.436     0.1274        155        640:  93%|█████████▎| 751/809 [14:45<01:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166     0.4359     0.1275        273        640:  93%|█████████▎| 753/809 [14:47<01:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.165     0.4361     0.1273        182        640:  93%|█████████▎| 755/809 [14:49<01:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166     0.4359     0.1274        360        640:  94%|█████████▎| 757/809 [14:51<01:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166     0.4358     0.1275        244        640:  94%|█████████▎| 758/809 [14:53<00:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.167     0.4357     0.1276        321        640:  94%|█████████▍| 760/809 [14:55<00:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.167     0.4357     0.1275        306        640:  94%|█████████▍| 762/809 [14:57<00:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.167     0.4356     0.1275        279        640:  94%|█████████▍| 763/809 [14:59<00:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.168     0.4354     0.1279        438        640:  95%|█████████▍| 765/809 [15:01<00:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.168     0.4353     0.1281        496        640:  95%|█████████▍| 767/809 [15:04<00:50,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.168     0.4353     0.1281        198        640:  95%|█████████▍| 768/809 [15:05<00:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.168     0.4355      0.128        316        640:  95%|█████████▌| 770/809 [15:07<00:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.168     0.4356     0.1279        170        640:  95%|█████████▌| 772/809 [15:09<00:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.167     0.4357     0.1279        323        640:  96%|█████████▌| 773/809 [15:11<00:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.167     0.4358     0.1278        147        640:  96%|█████████▌| 775/809 [15:13<00:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.167     0.4357      0.128        239        640:  96%|█████████▌| 777/809 [15:15<00:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.167     0.4357      0.128        312        640:  96%|█████████▋| 779/809 [15:18<00:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.167     0.4356     0.1279        324        640:  96%|█████████▋| 780/809 [15:19<00:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166     0.4359     0.1278         76        640:  97%|█████████▋| 782/809 [15:21<00:33,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.166      0.436     0.1277        306        640:  97%|█████████▋| 784/809 [15:24<00:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.165     0.4361     0.1277        230        640:  97%|█████████▋| 785/809 [15:25<00:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4365     0.1275        107        640:  97%|█████████▋| 787/809 [15:27<00:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4367     0.1274         92        640:  98%|█████████▊| 789/809 [15:30<00:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4367     0.1274        226        640:  98%|█████████▊| 790/809 [15:31<00:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4369     0.1273        219        640:  98%|█████████▊| 792/809 [15:33<00:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4368     0.1275        358        640:  98%|█████████▊| 794/809 [15:35<00:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4368     0.1274        164        640:  98%|█████████▊| 795/809 [15:37<00:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.164     0.4368     0.1277         99        640:  99%|█████████▊| 797/809 [15:39<00:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.163     0.4371     0.1276        138        640:  99%|█████████▉| 799/809 [15:41<00:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.162     0.4372     0.1275        165        640:  99%|█████████▉| 801/809 [15:44<00:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.163     0.4372     0.1275        279        640:  99%|█████████▉| 802/809 [15:45<00:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.163     0.4372     0.1276        519        640:  99%|█████████▉| 804/809 [15:47<00:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.163     0.4371     0.1276        406        640: 100%|█████████▉| 806/809 [15:50<00:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     82/300      14.1G      1.162     0.4372     0.1276        101        640: 100%|█████████▉| 807/809 [15:51<00:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   4%|▍         | 3/69 [00:00<00:16,  4.12it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  16%|█▌        | 11/69 [00:02<00:14,  4.08it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  28%|██▊       | 19/69 [00:04<00:12,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  39%|███▉      | 27/69 [00:06<00:10,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  51%|█████     | 35/69 [00:08<00:08,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  62%|██████▏   | 43/69 [00:10<00:06,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  74%|███████▍  | 51/69 [00:12<00:04,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  86%|████████▌ | 59/69 [00:14<00:02,  4.04it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:16<00:00,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:16<00:00,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4445     0.0926        207        640:   0%|          | 2/809 [00:02<15:13,  1.13s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.079     0.4615    0.09045        180        640:   0%|          | 3/809 [00:03<15:46,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.131      0.451     0.1078        189        640:   1%|          | 5/809 [00:05<15:50,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.094     0.4663     0.1049        157        640:   1%|          | 7/809 [00:08<15:57,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.126     0.4584     0.1107        329        640:   1%|          | 8/809 [00:09<16:02,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.072     0.4751     0.1066        277        640:   1%|          | 10/809 [00:11<15:56,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.109     0.4611     0.1145        362        640:   1%|▏         | 12/809 [00:14<15:51,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.113     0.4587     0.1138        292        640:   2%|▏         | 13/809 [00:15<15:48,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.137     0.4511     0.1244        512        640:   2%|▏         | 15/809 [00:17<15:43,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.123      0.451     0.1279        416        640:   2%|▏         | 17/809 [00:20<15:20,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.108     0.4534     0.1248        113        640:   2%|▏         | 18/809 [00:21<15:19,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.09     0.4569     0.1213         93        640:   2%|▏         | 20/809 [00:23<15:32,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.083     0.4607     0.1208        138        640:   3%|▎         | 22/809 [00:26<15:28,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.097     0.4577     0.1272        416        640:   3%|▎         | 23/809 [00:27<15:27,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.094     0.4592     0.1282        329        640:   3%|▎         | 25/809 [00:29<15:36,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.093       0.46     0.1271        376        640:   3%|▎         | 27/809 [00:32<15:54,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.089     0.4625     0.1265        233        640:   3%|▎         | 28/809 [00:33<15:32,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.077      0.466     0.1246        257        640:   4%|▎         | 30/809 [00:35<15:09,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.073     0.4665     0.1228        314        640:   4%|▍         | 32/809 [00:37<14:57,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.059      0.469     0.1209        102        640:   4%|▍         | 34/809 [00:40<14:52,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.055     0.4691     0.1203        245        640:   4%|▍         | 35/809 [00:41<15:08,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.047     0.4705     0.1187        220        640:   5%|▍         | 38/809 [00:44<15:00,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.047     0.4707     0.1177        365        640:   5%|▍         | 39/809 [00:45<14:55,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.051     0.4708     0.1212        378        640:   5%|▌         | 41/809 [00:48<14:49,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.049     0.4713       0.12        129        640:   5%|▌         | 43/809 [00:50<14:52,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.054     0.4709     0.1205        477        640:   5%|▌         | 44/809 [00:51<14:49,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.07     0.4682     0.1237        305        640:   6%|▌         | 46/809 [00:54<14:38,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.083     0.4665     0.1285        456        640:   6%|▌         | 48/809 [00:56<14:49,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.081     0.4682     0.1274        290        640:   6%|▌         | 50/809 [00:58<14:38,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.079     0.4693     0.1267        241        640:   6%|▋         | 51/809 [00:59<14:41,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.101     0.4626     0.1306        447        640:   7%|▋         | 53/809 [01:02<14:48,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.095     0.4627     0.1287        161        640:   7%|▋         | 55/809 [01:04<14:45,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.096     0.4617     0.1279        215        640:   7%|▋         | 57/809 [01:06<14:33,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.095      0.461     0.1273        283        640:   7%|▋         | 58/809 [01:08<14:28,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.099     0.4597     0.1268        221        640:   7%|▋         | 60/809 [01:10<14:46,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.103     0.4581     0.1261        164        640:   8%|▊         | 62/809 [01:12<14:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.109     0.4563     0.1261        185        640:   8%|▊         | 63/809 [01:13<14:33,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.118     0.4528     0.1262        298        640:   8%|▊         | 65/809 [01:16<15:05,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.127     0.4506     0.1265        235        640:   8%|▊         | 66/809 [01:17<15:10,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.129     0.4498     0.1269        155        640:   8%|▊         | 68/809 [01:20<15:17,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.131     0.4488     0.1261        260        640:   9%|▊         | 70/809 [01:22<14:45,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.13     0.4497     0.1253        171        640:   9%|▉         | 72/809 [01:24<14:28,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.131     0.4493     0.1251        272        640:   9%|▉         | 73/809 [01:26<14:29,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.129     0.4498     0.1247        136        640:   9%|▉         | 75/809 [01:28<14:34,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.129     0.4495     0.1241        291        640:  10%|▉         | 77/809 [01:30<14:30,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.132     0.4486     0.1246        420        640:  10%|▉         | 78/809 [01:32<14:22,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.132     0.4485     0.1248        275        640:  10%|▉         | 80/809 [01:34<14:04,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.134     0.4478     0.1244        219        640:  10%|█         | 82/809 [01:36<14:36,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.131     0.4488      0.124        129        640:  10%|█         | 83/809 [01:37<14:31,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.128     0.4497     0.1235        143        640:  11%|█         | 85/809 [01:40<14:11,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.131     0.4484     0.1241        300        640:  11%|█         | 87/809 [01:42<14:01,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.128     0.4494     0.1236        197        640:  11%|█         | 89/809 [01:44<13:54,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.125     0.4502     0.1234        159        640:  11%|█         | 90/809 [01:46<14:13,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.119     0.4517     0.1228        125        640:  11%|█▏        | 92/809 [01:48<13:59,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.117     0.4523     0.1226        302        640:  12%|█▏        | 94/809 [01:50<13:52,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.113     0.4542     0.1227        161        640:  12%|█▏        | 96/809 [01:53<13:46,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.111     0.4546     0.1224        116        640:  12%|█▏        | 97/809 [01:54<13:44,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.111     0.4542     0.1222        352        640:  12%|█▏        | 99/809 [01:56<13:51,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.117     0.4523     0.1229        546        640:  12%|█▏        | 101/809 [01:58<13:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.115     0.4532     0.1225         88        640:  13%|█▎        | 102/809 [02:00<13:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.114      0.454     0.1228        234        640:  13%|█▎        | 104/809 [02:02<13:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.116     0.4535     0.1226        271        640:  13%|█▎        | 106/809 [02:04<13:31,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.111     0.4547      0.122        145        640:  13%|█▎        | 108/809 [02:07<13:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.112      0.454      0.122        182        640:  13%|█▎        | 109/809 [02:08<13:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.115     0.4539     0.1221        285        640:  14%|█▎        | 111/809 [02:10<13:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.116      0.453     0.1223        114        640:  14%|█▍        | 113/809 [02:12<13:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.116     0.4532     0.1221        171        640:  14%|█▍        | 114/809 [02:14<13:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.114     0.4536     0.1217        183        640:  14%|█▍        | 116/809 [02:16<13:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.113     0.4542     0.1212        179        640:  15%|█▍        | 118/809 [02:18<13:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.119     0.4522     0.1218        246        640:  15%|█▍        | 120/809 [02:21<13:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.12     0.4517     0.1217        359        640:  15%|█▍        | 121/809 [02:22<13:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.124     0.4502     0.1224        241        640:  15%|█▌        | 123/809 [02:24<13:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.125     0.4494     0.1227        110        640:  15%|█▌        | 125/809 [02:27<13:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.128     0.4484     0.1231        264        640:  16%|█▌        | 126/809 [02:28<13:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.129      0.448     0.1228        240        640:  16%|█▌        | 128/809 [02:30<13:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.13     0.4475     0.1226        301        640:  16%|█▌        | 130/809 [02:32<13:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.135      0.446      0.124        479        640:  16%|█▌        | 131/809 [02:34<13:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.135     0.4457     0.1238        183        640:  16%|█▋        | 133/809 [02:36<13:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.136     0.4456     0.1241        405        640:  17%|█▋        | 135/809 [02:39<13:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.134     0.4464     0.1237        135        640:  17%|█▋        | 136/809 [02:40<13:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.133     0.4466     0.1237        239        640:  17%|█▋        | 138/809 [02:42<13:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.131      0.447     0.1237        197        640:  17%|█▋        | 140/809 [02:45<13:51,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.132     0.4466     0.1238        261        640:  17%|█▋        | 141/809 [02:46<13:34,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.131     0.4463     0.1237        259        640:  18%|█▊        | 143/809 [02:48<13:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.133     0.4457     0.1235        207        640:  18%|█▊        | 145/809 [02:50<12:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.134     0.4452     0.1236        152        640:  18%|���▊        | 147/809 [02:53<12:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.134     0.4456     0.1234        203        640:  18%|█▊        | 148/809 [02:54<13:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.136     0.4447     0.1249        685        640:  19%|█▊        | 150/809 [02:57<13:38,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.136     0.4447     0.1249        204        640:  19%|█▊        | 151/809 [02:58<13:29,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.139     0.4436      0.125        452        640:  19%|█▉        | 153/809 [03:00<13:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.141      0.443     0.1256        358        640:  19%|█▉        | 155/809 [03:02<13:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.141     0.4431     0.1257        303        640:  19%|█▉        | 156/809 [03:04<13:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.143     0.4426     0.1261        460        640:  20%|█▉        | 158/809 [03:06<13:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.143     0.4426     0.1259        212        640:  20%|█▉        | 160/809 [03:08<12:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.145     0.4418     0.1261        358        640:  20%|█▉        | 161/809 [03:10<13:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.149     0.4407     0.1264        432        640:  20%|██        | 163/809 [03:12<13:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.149     0.4404     0.1261        460        640:  20%|██        | 165/809 [03:14<12:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.148     0.4402     0.1259        191        640:  21%|██        | 167/809 [03:17<12:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.147     0.4403     0.1258        173        640:  21%|██        | 168/809 [03:18<12:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.146     0.4416     0.1254        156        640:  21%|██        | 170/809 [03:20<12:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.142     0.4424     0.1251        181        640:  21%|██▏       | 172/809 [03:23<12:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.143     0.4422     0.1252        252        640:  21%|██▏       | 173/809 [03:24<12:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.145     0.4414     0.1254        439        640:  22%|██▏       | 175/809 [03:26<12:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.145     0.4411     0.1254        265        640:  22%|██▏       | 177/809 [03:28<12:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.147     0.4406     0.1263        315        640:  22%|██▏       | 179/809 [03:31<12:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.147     0.4402     0.1261        171        640:  22%|██▏       | 181/809 [03:33<12:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.148     0.4397      0.126        298        640:  22%|██▏       | 182/809 [03:34<12:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.15     0.4392     0.1258        319        640:  23%|██▎       | 184/809 [03:37<12:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.152     0.4385     0.1258        368        640:  23%|██▎       | 186/809 [03:39<12:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.154      0.438     0.1263        323        640:  23%|██▎       | 188/809 [03:41<11:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4371     0.1267        453        640:  23%|██▎       | 189/809 [03:43<12:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4369     0.1267        249        640:  24%|██▎       | 191/809 [03:45<12:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4368     0.1269        273        640:  24%|██▍       | 193/809 [03:47<12:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4372     0.1266        129        640:  24%|██▍       | 194/809 [03:49<12:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153      0.437     0.1264        230        640:  24%|██▍       | 196/809 [03:51<12:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4359     0.1264        301        640:  24%|██▍       | 198/809 [03:53<11:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.158     0.4352     0.1274        366        640:  25%|██▍       | 199/809 [03:54<11:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.159     0.4347     0.1272        334        640:  25%|██▍       | 201/809 [03:57<11:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4337     0.1272        509        640:  25%|██▌       | 203/809 [03:59<11:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4332     0.1274        322        640:  25%|██▌       | 205/809 [04:01<11:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4336     0.1272        196        640:  25%|██▌       | 206/809 [04:03<11:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4335     0.1275        251        640:  26%|██▌       | 208/809 [04:05<11:31,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.16     0.4334     0.1272         57        640:  26%|██▌       | 210/809 [04:07<11:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161      0.433     0.1274        397        640:  26%|██▌       | 211/809 [04:08<11:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4327     0.1275        362        640:  26%|██▋       | 213/809 [04:11<11:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4325     0.1277        287        640:  27%|██▋       | 215/809 [04:13<11:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4325     0.1275        232        640:  27%|██▋       | 217/809 [04:15<11:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4329     0.1273        235        640:  27%|██▋       | 218/809 [04:17<11:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4323     0.1273        397        640:  27%|██▋       | 220/809 [04:19<11:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.158     0.4328     0.1267        150        640:  27%|██▋       | 222/809 [04:21<11:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.154     0.4337     0.1263        168        640:  28%|██▊       | 224/809 [04:23<11:14,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4336     0.1262        332        640:  28%|██▊       | 225/809 [04:25<11:14,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4337      0.126        256        640:  28%|██▊       | 227/809 [04:27<11:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4338     0.1258        196        640:  28%|██▊       | 229/809 [04:29<11:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4331     0.1266        565        640:  28%|██▊       | 230/809 [04:30<11:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4334     0.1265        189        640:  29%|██▊       | 232/809 [04:33<11:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.154     0.4335     0.1264        297        640:  29%|██▉       | 234/809 [04:35<11:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4337     0.1262        195        640:  29%|██▉       | 236/809 [04:37<11:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4337      0.126        164        640:  29%|██▉       | 237/809 [04:39<11:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.154     0.4333      0.126        403        640:  30%|██▉       | 239/809 [04:41<11:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4326     0.1263        558        640:  30%|██▉       | 241/809 [04:43<11:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4327     0.1261        350        640:  30%|██▉       | 242/809 [04:45<11:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4321     0.1267        169        640:  30%|███       | 244/809 [04:47<11:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4329     0.1264        132        640:  30%|███       | 246/809 [04:49<11:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4328     0.1265        243        640:  31%|███       | 247/809 [04:51<10:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.159     0.4322     0.1264        309        640:  31%|███       | 249/809 [04:53<10:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4325     0.1262        108        640:  31%|███       | 251/809 [04:55<10:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.158     0.4325     0.1261        375        640:  31%|███▏      | 253/809 [04:58<10:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4326     0.1259        292        640:  31%|███▏      | 254/809 [04:59<10:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4323     0.1258        297        640:  32%|███▏      | 256/809 [05:01<10:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4325     0.1256        281        640:  32%|███▏      | 258/809 [05:03<10:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4323     0.1258        376        640:  32%|███▏      | 259/809 [05:05<10:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4326     0.1255        206        640:  32%|███▏      | 261/809 [05:07<10:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4329     0.1254        173        640:  33%|███▎      | 263/809 [05:09<10:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4329     0.1252        247        640:  33%|███▎      | 264/809 [05:11<10:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4336      0.125        266        640:  33%|███▎      | 266/809 [05:13<10:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.154     0.4333     0.1253        153        640:  33%|███▎      | 268/809 [05:15<10:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.152     0.4339     0.1249        147        640:  33%|███▎      | 270/809 [05:18<10:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4333      0.126        440        640:  33%|███▎      | 271/809 [05:19<10:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4333     0.1258        344        640:  34%|███▎      | 273/809 [05:21<10:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4326     0.1264        562        640:  34%|███▍      | 275/809 [05:24<10:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4325     0.1262        248        640:  34%|███▍      | 276/809 [05:25<10:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4333     0.1258        180        640:  34%|███▍      | 278/809 [05:27<10:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.158     0.4325     0.1263        254        640:  35%|███▍      | 280/809 [05:29<10:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.158     0.4323     0.1263        457        640:  35%|███▍      | 281/809 [05:31<10:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4325     0.1264        350        640:  35%|███▍      | 283/809 [05:33<10:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4324     0.1262        242        640:  35%|███▌      | 285/809 [05:35<10:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.158     0.4321     0.1261        104        640:  35%|███▌      | 287/809 [05:38<10:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.158     0.4322      0.126        119        640:  36%|███▌      | 288/809 [05:39<10:00,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4324     0.1257        187        640:  36%|███▌      | 290/809 [05:41<09:53,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4331     0.1254        193        640:  36%|███▌      | 292/809 [05:43<09:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4328     0.1255        300        640:  36%|███▋      | 294/809 [05:46<09:50,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4327     0.1255        197        640:  36%|███▋      | 295/809 [05:47<09:47,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.154     0.4334     0.1252        232        640:  37%|███▋      | 297/809 [05:49<10:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4331     0.1256        382        640:  37%|███▋      | 298/809 [05:51<10:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4332     0.1255        143        640:  37%|███▋      | 300/809 [05:53<10:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4329     0.1255        322        640:  37%|███▋      | 302/809 [05:55<10:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4326     0.1254        285        640:  38%|███▊      | 304/809 [05:58<09:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.158     0.4323     0.1257        407        640:  38%|███▊      | 305/809 [05:59<09:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.159     0.4321     0.1258        374        640:  38%|███▊      | 307/809 [06:01<09:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.158     0.4325     0.1256        315        640:  38%|███▊      | 309/809 [06:03<09:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4326     0.1256        298        640:  38%|███▊      | 311/809 [06:06<09:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4327     0.1255        156        640:  39%|███▊      | 312/809 [06:07<09:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4328     0.1253        294        640:  39%|███▉      | 314/809 [06:09<09:28,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156      0.433     0.1252        140        640:  39%|███▉      | 316/809 [06:12<09:28,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156      0.433      0.125        267        640:  39%|███▉      | 317/809 [06:13<09:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4329     0.1248        169        640:  39%|███▉      | 319/809 [06:15<09:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4337     0.1246        174        640:  40%|███▉      | 321/809 [06:17<09:23,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.151     0.4342     0.1242        117        640:  40%|███▉      | 323/809 [06:20<09:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.151     0.4343     0.1242        306        640:  40%|████      | 325/809 [06:22<09:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.152     0.4344     0.1243        226        640:  40%|████      | 326/809 [06:23<09:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.151     0.4345     0.1242        269        640:  41%|████      | 328/809 [06:26<09:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.15     0.4346     0.1244        105        640:  41%|████      | 330/809 [06:28<09:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.151     0.4341     0.1244        342        640:  41%|████      | 332/809 [06:30<09:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.152     0.4338     0.1245        363        640:  41%|████      | 333/809 [06:31<09:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.154     0.4335     0.1244        310        640:  41%|████▏     | 335/809 [06:34<09:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4339     0.1242        232        640:  42%|████▏     | 337/809 [06:36<09:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.152     0.4341     0.1241        126        640:  42%|████▏     | 338/809 [06:37<09:02,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.152     0.4342     0.1239        114        640:  42%|████▏     | 340/809 [06:40<09:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.152     0.4343      0.124        229        640:  42%|████▏     | 342/809 [06:42<09:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4339     0.1242        318        640:  43%|████▎     | 344/809 [06:44<09:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4342     0.1241        180        640:  43%|████▎     | 345/809 [06:46<09:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4332     0.1266        501        640:  43%|████▎     | 347/809 [06:48<09:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4337     0.1264        161        640:  43%|████▎     | 349/809 [06:50<09:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4337     0.1264        347        640:  43%|████▎     | 350/809 [06:51<08:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.154     0.4341     0.1263        324        640:  44%|████▎     | 352/809 [06:54<08:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4346      0.126        270        640:  44%|████▍     | 354/809 [06:56<08:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4348      0.126        164        640:  44%|████▍     | 355/809 [06:57<08:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.152     0.4349     0.1259        169        640:  44%|████▍     | 357/809 [07:00<08:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.15     0.4354     0.1257         87        640:  44%|████▍     | 359/809 [07:02<08:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.15     0.4354     0.1256        308        640:  45%|████▍     | 361/809 [07:04<08:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.15     0.4354     0.1256        384        640:  45%|████▍     | 362/809 [07:05<08:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.149     0.4356     0.1255        248        640:  45%|████▍     | 364/809 [07:08<08:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.149     0.4357     0.1258        282        640:  45%|████▌     | 366/809 [07:10<08:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.148     0.4361     0.1257        129        640:  45%|████▌     | 367/809 [07:11<08:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.149     0.4358     0.1258        307        640:  46%|████▌     | 369/809 [07:14<08:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.149     0.4357     0.1256        185        640:  46%|████▌     | 371/809 [07:16<08:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.15     0.4353     0.1257        432        640:  46%|████▌     | 372/809 [07:17<08:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.151     0.4351     0.1256         89        640:  46%|████▌     | 374/809 [07:20<08:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.153     0.4346     0.1257        196        640:  46%|████▋     | 376/809 [07:22<08:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.155     0.4339     0.1259        314        640:  47%|████▋     | 378/809 [07:24<08:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4335     0.1265        408        640:  47%|████▋     | 379/809 [07:26<08:48,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4331     0.1265        314        640:  47%|████▋     | 381/809 [07:28<08:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4335     0.1263        155        640:  47%|████▋     | 383/809 [07:30<08:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.156     0.4335     0.1262        206        640:  47%|████▋     | 384/809 [07:32<08:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.157     0.4333     0.1261        295        640:  48%|████▊     | 386/809 [07:34<08:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.159     0.4327     0.1262        455        640:  48%|████▊     | 388/809 [07:36<08:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.159     0.4329     0.1262        124        640:  48%|████▊     | 390/809 [07:39<08:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.159      0.433     0.1261        210        640:  48%|████▊     | 391/809 [07:40<08:02,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4325     0.1265        282        640:  49%|████▊     | 393/809 [07:42<08:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4319     0.1266        314        640:  49%|████▉     | 395/809 [07:44<08:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4319     0.1265        172        640:  49%|████▉     | 396/809 [07:45<07:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4319     0.1265        160        640:  49%|████▉     | 398/809 [07:48<07:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163      0.432     0.1264        341        640:  49%|████▉     | 400/809 [07:50<07:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4323     0.1263        256        640:  50%|████▉     | 402/809 [07:52<07:48,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4322     0.1264        378        640:  50%|████▉     | 403/809 [07:54<07:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4317     0.1263        349        640:  50%|█████     | 405/809 [07:56<07:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4317     0.1262        135        640:  50%|█████     | 407/809 [07:58<07:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4312     0.1262        251        640:  51%|█████     | 409/809 [08:01<07:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4308     0.1264        492        640:  51%|█████     | 410/809 [08:02<07:40,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166      0.431     0.1264        276        640:  51%|█████     | 412/809 [08:04<07:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4311     0.1262        277        640:  51%|█████     | 414/809 [08:06<07:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165      0.431     0.1262        137        640:  51%|█████▏    | 415/809 [08:08<07:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4313      0.126        302        640:  52%|█████▏    | 417/809 [08:10<07:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162      0.432     0.1259        141        640:  52%|█████▏    | 419/809 [08:12<07:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4321     0.1258        191        640:  52%|█████▏    | 421/809 [08:15<07:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4321     0.1257        173        640:  52%|█████▏    | 422/809 [08:16<07:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4319     0.1257        143        640:  52%|█████▏    | 424/809 [08:18<07:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4321     0.1257        239        640:  53%|█████▎    | 426/809 [08:20<07:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4322     0.1255        211        640:  53%|█████▎    | 428/809 [08:23<07:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4324     0.1255        169        640:  53%|█████▎    | 429/809 [08:24<07:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.16     0.4326     0.1254        150        640:  53%|█████▎    | 431/809 [08:26<07:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.16     0.4327     0.1254        111        640:  54%|█████▎    | 433/809 [08:29<07:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.16     0.4328     0.1253        160        640:  54%|█████▎    | 434/809 [08:30<07:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.16     0.4329     0.1252        174        640:  54%|█████▍    | 436/809 [08:32<07:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4327     0.1251        222        640:  54%|█████▍    | 438/809 [08:34<07:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.16     0.4327     0.1251        111        640:  54%|█████▍    | 439/809 [08:36<07:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4326      0.125        147        640:  55%|█████▍    | 441/809 [08:38<07:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4325      0.125        221        640:  55%|█████▍    | 443/809 [08:41<07:28,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4325     0.1249        342        640:  55%|█████▍    | 444/809 [08:42<07:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4325     0.1249        172        640:  55%|█████▌    | 446/809 [08:44<07:24,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4324     0.1247        426        640:  55%|█████▌    | 448/809 [08:47<07:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4324     0.1247        275        640:  56%|█████▌    | 449/809 [08:48<07:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4325     0.1247        270        640:  56%|█████▌    | 451/809 [08:50<07:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4324     0.1247        332        640:  56%|█████▌    | 453/809 [08:53<07:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4325     0.1246        295        640:  56%|█████▌    | 454/809 [08:54<06:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4324     0.1245        154        640:  56%|█████▋    | 456/809 [08:56<06:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4317     0.1248        468        640:  57%|█████▋    | 458/809 [08:59<07:18,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4317     0.1248        227        640:  57%|█████▋    | 459/809 [09:00<07:15,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4319     0.1246        329        640:  57%|█████▋    | 461/809 [09:02<07:17,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4318     0.1245        192        640:  57%|█████▋    | 463/809 [09:05<06:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4319     0.1244        215        640:  57%|█████▋    | 464/809 [09:06<06:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4319     0.1243        143        640:  58%|█████▊    | 466/809 [09:08<06:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4316     0.1244        364        640:  58%|█████▊    | 468/809 [09:11<06:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4316     0.1243        228        640:  58%|█████▊    | 469/809 [09:12<06:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4312     0.1242        205        640:  58%|█████▊    | 471/809 [09:14<06:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4311     0.1241        172        640:  58%|█████▊    | 473/809 [09:17<06:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4309     0.1242        283        640:  59%|█████▊    | 474/809 [09:18<06:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164      0.431     0.1241        202        640:  59%|█████▉    | 476/809 [09:20<06:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4312      0.124        119        640:  59%|█████▉    | 478/809 [09:22<06:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4312     0.1239        411        640:  59%|█████▉    | 480/809 [09:25<06:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4312     0.1237        173        640:  60%|█████▉    | 482/809 [09:27<06:16,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162      0.431     0.1237        149        640:  60%|█████▉    | 484/809 [09:29<06:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4309     0.1237        246        640:  60%|█████▉    | 485/809 [09:30<06:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4307     0.1239        300        640:  60%|██████    | 487/809 [09:33<06:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4307     0.1238        227        640:  60%|██████    | 489/809 [09:35<06:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4307     0.1237        192        640:  61%|██████    | 490/809 [09:36<06:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4305     0.1238        318        640:  61%|██████    | 492/809 [09:39<06:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4304     0.1237        408        640:  61%|██████    | 494/809 [09:41<06:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4304     0.1236        276        640:  61%|██████    | 495/809 [09:42<06:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4303     0.1236        291        640:  61%|██████▏   | 497/809 [09:45<06:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4305     0.1235        191        640:  62%|██████▏   | 499/809 [09:47<06:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4305     0.1235        207        640:  62%|██████▏   | 501/809 [09:49<06:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4304     0.1236        421        640:  62%|██████▏   | 502/809 [09:50<05:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4302     0.1236        339        640:  62%|██████▏   | 504/809 [09:53<05:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4302     0.1237        157        640:  63%|██████▎   | 506/809 [09:55<05:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4302     0.1236        148        640:  63%|██████▎   | 507/809 [09:56<05:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4301     0.1235        325        640:  63%|██████▎   | 509/809 [09:59<05:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4303     0.1235        294        640:  63%|██████▎   | 511/809 [10:01<05:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4305     0.1234        156        640:  63%|██████▎   | 513/809 [10:03<05:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4303     0.1243        279        640:  64%|██████▎   | 514/809 [10:05<05:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4303     0.1243        176        640:  64%|██████▍   | 516/809 [10:07<05:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4303     0.1242        138        640:  64%|██████▍   | 518/809 [10:09<05:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163       0.43     0.1244        430        640:  64%|██████▍   | 519/809 [10:11<05:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164       0.43     0.1244        217        640:  64%|██████▍   | 521/809 [10:13<05:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4302     0.1243        262        640:  65%|██████▍   | 523/809 [10:15<05:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4303     0.1242        250        640:  65%|██████▍   | 524/809 [10:16<05:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4304     0.1242        309        640:  65%|██████▌   | 526/809 [10:19<05:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4302     0.1242        256        640:  65%|██████▌   | 528/809 [10:21<05:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4303     0.1242        155        640:  65%|██████▌   | 529/809 [10:22<05:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4299     0.1246        457        640:  66%|██████▌   | 531/809 [10:25<05:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4299     0.1245        153        640:  66%|██████▌   | 533/809 [10:27<05:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4302     0.1244        226        640:  66%|██████▌   | 535/809 [10:29<05:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4301     0.1243        285        640:  66%|██████▋   | 536/809 [10:31<05:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4303     0.1242        187        640:  67%|██████▋   | 538/809 [10:33<05:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.161     0.4303     0.1241        317        640:  67%|██████▋   | 540/809 [10:35<05:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4303     0.1241         87        640:  67%|██████▋   | 541/809 [10:37<05:23,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4297     0.1242        243        640:  67%|██████▋   | 543/809 [10:39<05:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4295     0.1241        155        640:  67%|██████▋   | 545/809 [10:41<05:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4293     0.1241        223        640:  67%|██████▋   | 546/809 [10:42<05:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4292      0.124        123        640:  68%|██████▊   | 548/809 [10:45<05:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166      0.429     0.1241        177        640:  68%|██████▊   | 550/809 [10:47<05:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166      0.429      0.124        257        640:  68%|██████▊   | 551/809 [10:48<05:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4288     0.1242        230        640:  68%|██████▊   | 553/809 [10:51<04:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4287     0.1241        292        640:  69%|██████▊   | 555/809 [10:53<04:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4286     0.1241        283        640:  69%|██████▉   | 557/809 [10:55<04:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4284     0.1241        305        640:  69%|██████▉   | 558/809 [10:57<04:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4282     0.1241        386        640:  69%|██████▉   | 560/809 [10:59<04:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.169      0.428      0.124        493        640:  69%|██████▉   | 562/809 [11:01<04:45,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.169     0.4279     0.1241        321        640:  70%|██████▉   | 564/809 [11:04<04:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168      0.428      0.124        119        640:  70%|██████▉   | 565/809 [11:05<04:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4281     0.1239        214        640:  70%|███████   | 567/809 [11:07<04:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.17     0.4277     0.1243        603        640:  70%|███████   | 569/809 [11:09<04:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G       1.17     0.4274     0.1247        610        640:  70%|███████   | 570/809 [11:11<04:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.171     0.4273     0.1248        373        640:  71%|███████   | 572/809 [11:13<04:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.169     0.4277     0.1247         65        640:  71%|███████   | 574/809 [11:15<04:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.169     0.4278     0.1247         94        640:  71%|███████   | 576/809 [11:18<04:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4278     0.1247        223        640:  71%|███████▏  | 577/809 [11:19<04:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4279     0.1248        399        640:  72%|███████▏  | 579/809 [11:21<04:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168      0.428     0.1248        361        640:  72%|███████▏  | 581/809 [11:24<04:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4281     0.1247        181        640:  72%|███████▏  | 582/809 [11:25<04:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4279     0.1247        137        640:  72%|███████▏  | 584/809 [11:27<04:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4278     0.1246        211        640:  72%|███████▏  | 586/809 [11:29<04:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4279     0.1246        157        640:  73%|███████▎  | 587/809 [11:31<04:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167      0.428     0.1245        330        640:  73%|███████▎  | 589/809 [11:33<04:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167      0.428     0.1245        269        640:  73%|███████▎  | 591/809 [11:35<04:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4281     0.1244        265        640:  73%|███████▎  | 593/809 [11:38<04:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4283     0.1244        238        640:  73%|███████▎  | 594/809 [11:39<04:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4284     0.1243        244        640:  74%|███████▎  | 596/809 [11:41<04:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4284     0.1244        429        640:  74%|███████▍  | 598/809 [11:43<04:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4284     0.1243        259        640:  74%|███████▍  | 599/809 [11:45<04:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4282     0.1243        293        640:  74%|███████▍  | 601/809 [11:47<03:59,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4283     0.1243        382        640:  75%|███████▍  | 603/809 [11:49<04:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4283     0.1244        290        640:  75%|███████▍  | 605/809 [11:52<04:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4284     0.1243        308        640:  75%|███████▍  | 606/809 [11:53<03:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4286     0.1243        286        640:  75%|███████▌  | 608/809 [11:55<03:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4285     0.1243        395        640:  75%|███████▌  | 610/809 [11:58<03:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4287     0.1243        135        640:  76%|███████▌  | 611/809 [11:59<03:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165      0.429     0.1241        163        640:  76%|███████▌  | 614/809 [12:02<03:44,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165      0.429     0.1242        139        640:  76%|███████▌  | 615/809 [12:03<03:41,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4292     0.1241        289        640:  76%|███████▋  | 617/809 [12:06<03:40,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4294      0.124        229        640:  77%|███████▋  | 619/809 [12:08<03:39,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4294      0.124        249        640:  77%|███████▋  | 620/809 [12:09<03:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4293     0.1241        286        640:  77%|███████▋  | 622/809 [12:12<03:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4292     0.1242        404        640:  77%|███████▋  | 624/809 [12:14<03:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4291     0.1243        309        640:  77%|███████▋  | 626/809 [12:16<03:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4292     0.1243        163        640:  78%|███████▊  | 627/809 [12:18<03:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4292     0.1243        197        640:  78%|███████▊  | 629/809 [12:20<03:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4291     0.1248        126        640:  78%|███████▊  | 631/809 [12:22<03:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4291     0.1248        164        640:  78%|███████▊  | 632/809 [12:23<03:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4292     0.1247        205        640:  78%|███████▊  | 634/809 [12:26<03:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4292     0.1247        201        640:  79%|███████▊  | 636/809 [12:28<03:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4292     0.1246        116        640:  79%|███████▊  | 637/809 [12:29<03:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4296     0.1245        121        640:  79%|███████▉  | 639/809 [12:32<03:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4296     0.1245        413        640:  79%|███████▉  | 641/809 [12:34<03:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4297     0.1245        162        640:  79%|███████▉  | 642/809 [12:35<03:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4296     0.1246        189        640:  80%|███████▉  | 644/809 [12:38<03:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4298     0.1245        106        640:  80%|███████▉  | 646/809 [12:40<03:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4297     0.1245        321        640:  80%|████████  | 648/809 [12:42<03:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4297     0.1245        151        640:  80%|████████  | 649/809 [12:43<03:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4299     0.1245        213        640:  80%|████████  | 651/809 [12:46<03:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4301     0.1245        197        640:  81%|████████  | 653/809 [12:48<03:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4301     0.1244        124        640:  81%|████████  | 654/809 [12:49<03:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4301     0.1244        221        640:  81%|████████  | 656/809 [12:52<02:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4303     0.1244        270        640:  81%|████████▏ | 658/809 [12:54<02:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4304     0.1244        179        640:  81%|████████▏ | 659/809 [12:55<02:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4307     0.1243        129        640:  82%|████████▏ | 661/809 [12:58<02:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4306     0.1243        163        640:  82%|████████▏ | 663/809 [13:00<02:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4305     0.1243        203        640:  82%|████████▏ | 665/809 [13:02<02:47,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4302     0.1245        446        640:  82%|████████▏ | 666/809 [13:03<02:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4303     0.1247        469        640:  83%|████████▎ | 668/809 [13:06<02:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.162     0.4306     0.1246        148        640:  83%|████████▎ | 670/809 [13:08<02:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4306     0.1252        299        640:  83%|████████▎ | 672/809 [13:10<02:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4306     0.1252        309        640:  83%|████████▎ | 673/809 [13:12<02:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4307     0.1255        298        640:  83%|████████▎ | 675/809 [13:14<02:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4304     0.1256        342        640:  84%|████████▎ | 677/809 [13:16<02:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4303     0.1256        320        640:  84%|████████▍ | 678/809 [13:18<02:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4302     0.1255        273        640:  84%|████████▍ | 680/809 [13:20<02:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4301     0.1256        209        640:  84%|████████▍ | 682/809 [13:22<02:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4301     0.1255        175        640:  85%|████████▍ | 684/809 [13:25<02:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4302     0.1255        145        640:  85%|████████▍ | 685/809 [13:26<02:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4303     0.1254        253        640:  85%|████████▍ | 687/809 [13:28<02:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4301     0.1255        289        640:  85%|████████▌ | 689/809 [13:30<02:18,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4301     0.1255        228        640:  85%|████████▌ | 690/809 [13:32<02:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164       0.43     0.1254        170        640:  86%|████████▌ | 692/809 [13:34<02:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4299     0.1254        355        640:  86%|████████▌ | 694/809 [13:36<02:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4299     0.1254        198        640:  86%|████████▌ | 695/809 [13:38<02:24,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4299     0.1253        257        640:  86%|████████▌ | 697/809 [13:40<02:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4296     0.1252        200        640:  86%|████████▋ | 699/809 [13:42<02:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4298     0.1252        167        640:  87%|████████▋ | 700/809 [13:43<02:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4301     0.1251        288        640:  87%|████████▋ | 702/809 [13:46<02:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4303      0.125        301        640:  87%|████████▋ | 704/809 [13:48<02:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4306     0.1249        320        640:  87%|████████▋ | 706/809 [13:51<02:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4307     0.1248        166        640:  87%|████████▋ | 707/809 [13:52<02:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4306     0.1249        160        640:  88%|████████▊ | 709/809 [13:54<01:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4305     0.1249        223        640:  88%|████████▊ | 711/809 [13:56<01:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4305     0.1249        321        640:  88%|████████▊ | 712/809 [13:58<01:52,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4303     0.1249        342        640:  88%|████████▊ | 714/809 [14:00<01:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4301     0.1249        343        640:  89%|████████▊ | 716/809 [14:03<01:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4302     0.1249        232        640:  89%|████████▊ | 717/809 [14:04<01:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4302     0.1248        255        640:  89%|████████▉ | 719/809 [14:06<01:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4303     0.1247        292        640:  89%|████████▉ | 721/809 [14:08<01:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4303     0.1248        219        640:  89%|████████▉ | 722/809 [14:09<01:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4301      0.125        357        640:  89%|████████▉ | 724/809 [14:12<01:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4299     0.1251        224        640:  90%|████████▉ | 726/809 [14:14<01:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4297     0.1251        202        640:  90%|████████▉ | 728/809 [14:17<01:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4294     0.1254        233        640:  90%|█████████ | 729/809 [14:18<01:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.169     0.4292     0.1255        191        640:  90%|█████████ | 731/809 [14:20<01:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4294     0.1255        300        640:  91%|█████████ | 733/809 [14:23<01:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4296     0.1254        136        640:  91%|█████████ | 734/809 [14:24<01:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.168     0.4295     0.1254        204        640:  91%|█████████ | 736/809 [14:26<01:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4299     0.1253        149        640:  91%|█████████ | 738/809 [14:28<01:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4297     0.1254        329        640:  91%|█��███████▏| 740/809 [14:31<01:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4299     0.1253        140        640:  92%|█████████▏| 741/809 [14:32<01:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4299     0.1253        169        640:  92%|█████████▏| 743/809 [14:34<01:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166       0.43     0.1252        111        640:  92%|█████████▏| 745/809 [14:37<01:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4301     0.1252        176        640:  92%|█████████▏| 746/809 [14:38<01:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4299     0.1254        279        640:  92%|█████████▏| 748/809 [14:40<01:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167       0.43     0.1253        350        640:  93%|█████████▎| 750/809 [14:42<01:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167       0.43     0.1253        181        640:  93%|█████████▎| 752/809 [14:45<01:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167       0.43     0.1253        188        640:  93%|█████████▎| 753/809 [14:46<01:04,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4301     0.1253        175        640:  93%|█████████▎| 755/809 [14:48<01:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4299     0.1253        293        640:  94%|████████���▎| 757/809 [14:51<01:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4298     0.1254        470        640:  94%|█████████▍| 759/809 [14:53<00:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4297     0.1254        252        640:  94%|█████████▍| 761/809 [14:55<00:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4297     0.1253        168        640:  94%|█████████▍| 762/809 [14:56<00:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4299     0.1253         66        640:  94%|█████████▍| 764/809 [14:59<00:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4298     0.1253         88        640:  95%|█████████▍| 766/809 [15:01<00:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.167     0.4299     0.1253        118        640:  95%|█████████▍| 767/809 [15:02<00:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4301     0.1252        331        640:  95%|█████████▌| 769/809 [15:05<00:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4301     0.1252        351        640:  95%|█████████▌| 771/809 [15:07<00:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4301     0.1253        418        640:  95%|█████████▌| 772/809 [15:08<00:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4304     0.1252        238        640:  96%|█████████▌| 774/809 [15:11<00:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.166     0.4304     0.1253        288        640:  96%|█████████▌| 776/809 [15:13<00:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4305     0.1253        131        640:  96%|█████████▌| 778/809 [15:15<00:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4306     0.1252        186        640:  96%|█████████▋| 779/809 [15:17<00:36,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4307     0.1251        171        640:  97%|█████████▋| 781/809 [15:19<00:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4307     0.1252        252        640:  97%|█████████▋| 783/809 [15:21<00:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4307     0.1251        203        640:  97%|█████████▋| 784/809 [15:22<00:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165      0.431     0.1251        194        640:  97%|█████████▋| 786/809 [15:25<00:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165      0.431     0.1251        315        640:  97%|█████████▋| 788/809 [15:27<00:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4309      0.125        220        640:  98%|█████████▊| 789/809 [15:28<00:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165      0.431      0.125        151        640:  98%|█████████▊| 791/809 [15:31<00:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4313      0.125        223        640:  98%|█████████▊| 793/809 [15:33<00:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.165     0.4312      0.125        123        640:  98%|█████████▊| 794/809 [15:34<00:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4314     0.1249        289        640:  98%|█████████▊| 796/809 [15:37<00:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4314     0.1249        505        640:  99%|█████████▊| 798/809 [15:39<00:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4315     0.1248        162        640:  99%|█████████▉| 800/809 [15:41<00:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4316     0.1247        171        640:  99%|█████████▉| 801/809 [15:42<00:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4318     0.1246        138        640:  99%|█████████▉| 803/809 [15:45<00:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163      0.432     0.1246        140        640: 100%|█████████▉| 805/809 [15:47<00:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.163     0.4319     0.1246        177        640: 100%|█████████▉| 806/809 [15:48<00:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     83/300      13.7G      1.164     0.4317     0.1249        499        640: 100%|█████████▉| 808/809 [15:51<00:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   9%|▊         | 6/69 [00:01<00:15,  4.20it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  20%|██        | 14/69 [00:03<00:14,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  32%|███▏      | 22/69 [00:05<00:11,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  43%|████▎     | 30/69 [00:07<00:09,  4.09it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  55%|█████▌    | 38/69 [00:09<00:07,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  67%|██████▋   | 46/69 [00:11<00:06,  3.74it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  78%|███████▊  | 54/69 [00:13<00:03,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  90%|████████▉ | 62/69 [00:15<00:01,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  90%|████████▉ | 62/69 [00:15<00:01,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.01it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.2G       1.52     0.3291     0.1815        585        640:   0%|          | 1/809 [00:01<15:46,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.404     0.3645     0.1465        135        640:   0%|          | 2/809 [00:02<16:10,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.174     0.4386     0.1235        246        640:   0%|          | 4/809 [00:04<15:42,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.083     0.4663      0.116        149        640:   1%|          | 6/809 [00:07<15:37,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.071     0.4732     0.1159        308        640:   1%|          | 7/809 [00:08<15:26,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.044     0.4855     0.1136        176        640:   1%|          | 9/809 [00:10<15:26,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.057     0.4779     0.1147        365        640:   1%|▏         | 11/809 [00:12<15:31,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.118     0.4564     0.1255        524        640:   2%|▏         | 13/809 [00:15<15:38,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.128     0.4516     0.1241        335        640:   2%|▏         | 14/809 [00:16<15:34,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.149     0.4473     0.1624        183        640:   2%|▏         | 16/809 [00:18<15:39,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.138     0.4527     0.1638        134        640:   2%|▏         | 18/809 [00:21<15:33,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.166     0.4443     0.1752        405        640:   2%|▏         | 19/809 [00:22<16:01,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.169     0.4374     0.1695        307        640:   3%|▎         | 21/809 [00:24<15:28,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.148     0.4418     0.1626        245        640:   3%|▎         | 23/809 [00:27<15:19,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.141     0.4439     0.1598        199        640:   3%|▎         | 24/809 [00:28<15:21,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.152     0.4388     0.1564        281        640:   3%|▎         | 26/809 [00:30<15:34,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.167     0.4334      0.159        480        640:   3%|▎         | 28/809 [00:33<15:29,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4324     0.1581        306        640:   4%|▎         | 29/809 [00:34<15:33,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.175     0.4305     0.1557        231        640:   4%|▍         | 31/809 [00:36<15:23,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178      0.428     0.1544        231        640:   4%|▍         | 33/809 [00:38<15:10,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4324     0.1529        201        640:   4%|▍         | 35/809 [00:41<15:08,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.156     0.4347     0.1532         67        640:   4%|▍         | 36/809 [00:42<15:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.171     0.4341     0.1558         99        640:   5%|▍         | 38/809 [00:44<14:57,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182      0.431     0.1555        217        640:   5%|▍         | 40/809 [00:47<15:08,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.191     0.4284     0.1556        367        640:   5%|▌         | 41/809 [00:48<15:03,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4318     0.1539        163        640:   5%|▌         | 43/809 [00:50<14:55,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.186     0.4335     0.1525        143        640:   6%|▌         | 45/809 [00:52<14:47,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4366     0.1537        291        640:   6%|▌         | 47/809 [00:55<14:34,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4376     0.1527        135        640:   6%|▌         | 48/809 [00:56<14:43,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4379     0.1512        197        640:   6%|▌         | 50/809 [00:58<14:45,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4345     0.1517        182        640:   6%|▋         | 52/809 [01:00<14:25,  1.14s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.201      0.433     0.1533        310        640:   7%|▋         | 53/809 [01:02<15:07,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.217     0.4284     0.1544        292        640:   7%|▋         | 55/809 [01:04<14:51,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.228     0.4244     0.1545        199        640:   7%|▋         | 57/809 [01:06<14:36,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.246     0.4198     0.1578        353        640:   7%|▋         | 59/809 [01:09<14:37,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.254     0.4183     0.1598        427        640:   7%|▋         | 60/809 [01:10<14:38,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.261     0.4158     0.1623        371        640:   8%|▊         | 62/809 [01:12<14:36,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.255     0.4183     0.1607        227        640:   8%|▊         | 64/809 [01:15<14:57,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.257     0.4178     0.1607        166        640:   8%|▊         | 65/809 [01:16<14:50,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.262     0.4163     0.1605        238        640:   8%|▊         | 67/809 [01:18<14:57,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.254     0.4188     0.1593         95        640:   9%|▊         | 69/809 [01:21<14:51,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.258     0.4181     0.1606        413        640:   9%|▊         | 70/809 [01:22<14:40,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.259     0.4171     0.1601        207        640:   9%|▉         | 72/809 [01:24<14:23,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.256     0.4176     0.1589        319        640:   9%|▉         | 74/809 [01:27<14:21,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.253     0.4183     0.1581        316        640:   9%|▉         | 76/809 [01:29<14:12,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.25     0.4191     0.1573        191        640:  10%|▉         | 77/809 [01:30<14:10,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.247     0.4205     0.1597        158        640:  10%|▉         | 79/809 [01:32<14:20,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.244     0.4214      0.159        246        640:  10%|█         | 81/809 [01:35<14:07,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.247     0.4206     0.1609        571        640:  10%|█         | 82/809 [01:36<14:33,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.245      0.421     0.1603        270        640:  10%|█         | 84/809 [01:38<14:08,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.246     0.4209     0.1617        498        640:  11%|█         | 86/809 [01:41<14:22,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.25     0.4195     0.1629        330        640:  11%|█         | 87/809 [01:42<14:48,  1.23s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.246     0.4198      0.162        218        640:  11%|█         | 89/809 [01:44<14:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.245     0.4195     0.1614        290        640:  11%|█         | 91/809 [01:47<14:34,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.248     0.4183     0.1617        403        640:  11%|█▏        | 92/809 [01:48<14:20,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.252     0.4168     0.1632        340        640:  12%|█▏        | 95/809 [01:52<14:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.253      0.416     0.1629        293        640:  12%|█▏        | 96/809 [01:53<13:56,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.252     0.4152     0.1621        254        640:  12%|█▏        | 98/809 [01:55<13:56,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.25      0.416     0.1609        264        640:  12%|█▏        | 100/809 [01:57<13:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.251     0.4152     0.1607        348        640:  12%|█▏        | 101/809 [01:59<13:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.251     0.4148     0.1599        357        640:  13%|█▎        | 103/809 [02:01<13:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.247      0.416     0.1592        285        640:  13%|█▎        | 105/809 [02:03<13:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.247     0.4161     0.1585        294        640:  13%|█▎        | 107/809 [02:06<13:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.243     0.4169     0.1579        118        640:  13%|█▎        | 108/809 [02:07<13:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.246      0.416     0.1582        331        640:  14%|█▎        | 110/809 [02:09<13:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.248     0.4153     0.1581        326        640:  14%|█▍        | 112/809 [02:11<13:23,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.249      0.415     0.1581        238        640:  14%|█▍        | 113/809 [02:12<13:20,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.248     0.4153     0.1577        237        640:  14%|█▍        | 115/809 [02:15<13:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.242     0.4164     0.1567        205        640:  14%|█▍        | 117/809 [02:17<13:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.242     0.4162     0.1563        291        640:  15%|█▍        | 119/809 [02:20<13:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.245     0.4151     0.1564        430        640:  15%|█▍        | 120/809 [02:21<13:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.243     0.4159     0.1559        336        640:  15%|█▌        | 122/809 [02:23<13:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.245     0.4153     0.1568        489        640:  15%|█▌        | 124/809 [02:25<13:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.243     0.4158     0.1562        170        640:  15%|█▌        | 125/809 [02:27<13:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.243     0.4162     0.1563        454        640:  16%|█▌        | 127/809 [02:29<13:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.245     0.4158     0.1567        291        640:  16%|█▌        | 129/809 [02:31<13:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.243     0.4166     0.1557        248        640:  16%|█▌        | 130/809 [02:34<13:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.242     0.4169     0.1554        192        640:  16%|█▋        | 132/809 [02:35<13:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.242     0.4169     0.1547        178        640:  17%|█▋        | 134/809 [02:37<13:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.238     0.4182     0.1544        307        640:  17%|█▋        | 136/809 [02:40<13:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.238     0.4182     0.1539        205        640:  17%|█▋        | 137/809 [02:41<13:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.236     0.4187      0.153        242        640:  17%|█▋        | 139/809 [02:43<13:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.236     0.4184     0.1526        414        640:  17%|█▋        | 141/809 [02:45<13:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.237      0.418     0.1525        398        640:  18%|█▊        | 142/809 [02:47<12:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.235     0.4186      0.152        203        640:  18%|█▊        | 144/809 [02:49<12:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.233     0.4189     0.1512        280        640:  18%|█▊        | 146/809 [02:51<13:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.231     0.4195     0.1507        481        640:  18%|█▊        | 148/809 [02:54<13:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.234     0.4189     0.1519        899        640:  18%|█▊        | 149/809 [02:55<13:26,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.232     0.4194     0.1518        364        640:  19%|█▊        | 151/809 [02:57<13:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.229     0.4206     0.1509        252        640:  19%|█▉        | 153/809 [03:00<12:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.228     0.4211     0.1507        234        640:  19%|█▉        | 154/809 [03:01<12:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.227     0.4213     0.1502        443        640:  19%|█▉        | 156/809 [03:03<12:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.223     0.4222     0.1495        132        640:  20%|█▉        | 158/809 [03:05<12:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.224      0.422     0.1494        296        640:  20%|█▉        | 159/809 [03:07<12:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.223     0.4219     0.1492        196        640:  20%|█▉        | 161/809 [03:09<12:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.224     0.4216     0.1499        329        640:  20%|██        | 163/809 [03:11<12:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.226      0.421     0.1504        416        640:  20%|██        | 165/809 [03:14<12:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.225     0.4211     0.1501        179        640:  21%|██        | 166/809 [03:15<12:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.224     0.4211     0.1503        464        640:  21%|██        | 168/809 [03:17<12:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.223     0.4214     0.1499        190        640:  21%|██        | 170/809 [03:20<12:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.225     0.4205       0.15        375        640:  21%|██        | 171/809 [03:21<12:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.229     0.4191     0.1511        469        640:  21%|██▏       | 173/809 [03:23<12:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.229     0.4196      0.151        137        640:  22%|██▏       | 175/809 [03:25<12:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.232     0.4192     0.1517        328        640:  22%|██▏       | 177/809 [03:28<12:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.232      0.419     0.1516        289        640:  22%|██▏       | 178/809 [03:29<12:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.231     0.4194     0.1511        146        640:  22%|██▏       | 180/809 [03:31<12:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.227     0.4205     0.1506        297        640:  22%|██▏       | 182/809 [03:34<12:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.225     0.4214     0.1503        155        640:  23%|██▎       | 183/809 [03:35<12:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.222     0.4223     0.1497        284        640:  23%|██▎       | 185/809 [03:37<12:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.221      0.423     0.1494        241        640:  23%|██▎       | 187/809 [03:40<12:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.22     0.4232     0.1491        294        640:  23%|██▎       | 188/809 [03:41<12:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.22     0.4232      0.149        416        640:  23%|██▎       | 190/809 [03:43<12:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.222      0.423     0.1491        299        640:  24%|██▎       | 192/809 [03:46<12:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.22     0.4236      0.149        245        640:  24%|██▍       | 193/809 [03:47<12:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.221     0.4237      0.149        367        640:  24%|██▍       | 195/809 [03:49<12:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.219     0.4248     0.1484        241        640:  24%|██▍       | 197/809 [03:51<11:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.218     0.4253     0.1479        243        640:  25%|██▍       | 199/809 [03:54<11:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.217     0.4255     0.1478        238        640:  25%|██▍       | 200/809 [03:55<11:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.215     0.4263     0.1474        197        640:  25%|██▍       | 202/809 [03:57<12:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.213     0.4271     0.1467        168        640:  25%|██▌       | 204/809 [04:00<11:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.211     0.4275     0.1465        185        640:  25%|██▌       | 205/809 [04:01<11:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.21     0.4281     0.1464        192        640:  26%|██▌       | 207/809 [04:03<11:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.21     0.4281     0.1461        179        640:  26%|██▌       | 209/809 [04:06<11:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.208     0.4284     0.1456        316        640:  26%|██▌       | 211/809 [04:08<11:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.207     0.4287     0.1454        108        640:  26%|██▌       | 212/809 [04:09<11:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.204     0.4296     0.1449        223        640:  26%|██▋       | 214/809 [04:11<11:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.202     0.4307     0.1443        202        640:  27%|██▋       | 216/809 [04:14<11:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.202     0.4306     0.1441        331        640:  27%|██▋       | 217/809 [04:15<11:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.202     0.4304     0.1438        363        640:  27%|██▋       | 219/809 [04:17<11:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.201      0.431      0.144        428        640:  27%|██▋       | 221/809 [04:20<11:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.199     0.4316     0.1435        316        640:  28%|██▊       | 223/809 [04:22<11:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.197     0.4323     0.1431        249        640:  28%|██▊       | 224/809 [04:23<11:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4328     0.1429        349        640:  28%|██▊       | 226/809 [04:25<11:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.197     0.4329     0.1428        257        640:  28%|██▊       | 228/809 [04:28<11:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.197     0.4327     0.1426        223        640:  28%|██▊       | 229/809 [04:29<11:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.197      0.433     0.1424        319        640:  29%|██▊       | 231/809 [04:31<11:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4336     0.1421        203        640:  29%|██▉       | 233/809 [04:34<11:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4332     0.1418        320        640:  29%|██▉       | 235/809 [04:36<11:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4333     0.1416        374        640:  29%|██▉       | 236/809 [04:37<11:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4338     0.1412        229        640:  29%|██▉       | 238/809 [04:40<11:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.191     0.4345     0.1408        175        640:  30%|██▉       | 240/809 [04:42<11:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4342     0.1423        167        640:  30%|██▉       | 242/809 [04:44<11:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.191     0.4351      0.142        164        640:  30%|███       | 244/809 [04:47<11:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.191     0.4352      0.142        440        640:  30%|███       | 245/809 [04:48<10:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.189     0.4359     0.1415        362        640:  31%|███       | 247/809 [04:50<10:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4369      0.141        316        640:  31%|███       | 249/809 [04:52<11:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4366      0.141        288        640:  31%|███       | 250/809 [04:54<11:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4365     0.1408        337        640:  31%|███       | 252/809 [04:56<10:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4371     0.1405        190        640:  31%|███▏      | 254/809 [04:58<10:37,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.189      0.437     0.1408        477        640:  32%|███▏      | 256/809 [05:00<10:34,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.189     0.4369     0.1407        294        640:  32%|███▏      | 257/809 [05:02<10:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4375     0.1402         87        640:  32%|███▏      | 259/809 [05:04<10:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4375     0.1399        183        640:  32%|███▏      | 261/809 [05:06<10:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4374       0.14        268        640:  32%|███▏      | 262/809 [05:08<10:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.189     0.4371     0.1398        309        640:  33%|███▎      | 264/809 [05:10<10:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4379     0.1395        293        640:  33%|███▎      | 266/809 [05:12<10:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.186     0.4381     0.1392        308        640:  33%|███▎      | 268/809 [05:15<10:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.186     0.4381      0.139        200        640:  33%|███▎      | 269/809 [05:16<10:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4386     0.1387         93        640:  33%|███▎      | 271/809 [05:18<10:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4394     0.1382        166        640:  34%|███▎      | 273/809 [05:20<10:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4398     0.1379        182        640:  34%|███▍      | 274/809 [05:22<10:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181       0.44     0.1378        194        640:  34%|███▍      | 276/809 [05:24<10:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4401     0.1376        245        640:  34%|███▍      | 278/809 [05:26<10:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4401     0.1372        314        640:  35%|███▍      | 280/809 [05:29<10:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4402     0.1371        232        640:  35%|███▍      | 281/809 [05:30<10:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4398     0.1373        213        640:  35%|███▍      | 283/809 [05:32<10:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4393     0.1376        244        640:  35%|███▌      | 285/809 [05:35<10:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.183     0.4393     0.1376        374        640:  35%|███▌      | 286/809 [05:36<10:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.183     0.4393     0.1377        645        640:  36%|███▌      | 288/809 [05:38<10:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4397     0.1374        214        640:  36%|███▌      | 290/809 [05:41<10:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4399     0.1372        234        640:  36%|███▌      | 291/809 [05:42<10:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4397     0.1369        208        640:  36%|███▌      | 293/809 [05:44<10:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4399     0.1369        183        640:  36%|███▋      | 295/809 [05:46<09:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4397     0.1369        265        640:  37%|███▋      | 296/809 [05:48<09:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4398     0.1367        239        640:  37%|███▋      | 298/809 [05:50<10:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4396     0.1368        342        640:  37%|███▋      | 300/809 [05:53<10:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4398     0.1366        149        640:  37%|███▋      | 301/809 [05:54<10:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.177     0.4403     0.1362        261        640:  37%|███▋      | 303/809 [05:56<10:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.176     0.4406     0.1363        219        640:  38%|███▊      | 305/809 [05:58<09:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.176     0.4406     0.1362        276        640:  38%|███▊      | 306/809 [06:00<10:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.174     0.4408     0.1359        230        640:  38%|███▊      | 308/809 [06:02<10:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.174      0.441      0.136        126        640:  38%|███▊      | 310/809 [06:04<09:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.172     0.4415     0.1357        239        640:  39%|███▊      | 312/809 [06:07<09:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.171     0.4418     0.1355        199        640:  39%|███▊      | 313/809 [06:08<09:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4419     0.1354        274        640:  39%|███▉      | 315/809 [06:10<09:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17      0.442     0.1352        227        640:  39%|███▉      | 317/809 [06:13<09:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4419      0.135        240        640:  39%|███▉      | 318/809 [06:14<09:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.171     0.4415     0.1353        386        640:  40%|███▉      | 320/809 [06:16<09:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.172     0.4413     0.1356        238        640:  40%|███▉      | 322/809 [06:19<09:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.171     0.4415     0.1355        174        640:  40%|████      | 324/809 [06:21<09:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.171     0.4413     0.1356        495        640:  40%|████      | 325/809 [06:22<09:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.171     0.4414     0.1352        272        640:  40%|████      | 327/809 [06:24<09:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4415     0.1351        172        640:  41%|████      | 328/809 [06:26<09:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.168     0.4421     0.1348        149        640:  41%|████      | 330/809 [06:28<09:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.169     0.4419     0.1348        281        640:  41%|████      | 332/809 [06:31<09:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.168     0.4421     0.1346        247        640:  41%|████▏     | 334/809 [06:33<09:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.168     0.4424     0.1345        284        640:  41%|████▏     | 335/809 [06:34<09:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.168     0.4424     0.1344        293        640:  42%|████▏     | 337/809 [06:36<09:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4417     0.1344        356        640:  42%|████▏     | 339/809 [06:39<09:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.169     0.4417     0.1342        280        640:  42%|████▏     | 340/809 [06:40<09:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.169     0.4419      0.134        308        640:  42%|████▏     | 342/809 [06:42<09:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.167     0.4422     0.1338        271        640:  43%|████▎     | 344/809 [06:45<09:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.168      0.442     0.1344        562        640:  43%|████▎     | 345/809 [06:46<09:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.169     0.4418     0.1344        203        640:  43%|████▎     | 347/809 [06:48<09:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4416     0.1343        408        640:  43%|████▎     | 349/809 [06:51<08:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.169     0.4419     0.1342        215        640:  43%|████▎     | 351/809 [06:53<08:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4416     0.1343        359        640:  44%|████▎     | 352/809 [06:54<08:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.169      0.442      0.134        159        640:  44%|████▍     | 354/809 [06:56<08:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.168     0.4421     0.1338         78        640:  44%|████▍     | 356/809 [06:59<08:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4417     0.1346        657        640:  44%|████▍     | 357/809 [07:00<09:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.168     0.4423     0.1343        246        640:  44%|████▍     | 359/809 [07:02<08:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.171     0.4414     0.1344        626        640:  45%|████▍     | 361/809 [07:05<08:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4415     0.1342        181        640:  45%|████▍     | 362/809 [07:06<08:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.169     0.4416     0.1339        258        640:  45%|████▍     | 364/809 [07:08<08:54,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4413      0.134        300        640:  45%|████▌     | 366/809 [07:11<08:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.172     0.4408     0.1344        788        640:  45%|████▌     | 367/809 [07:12<09:00,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.175     0.4401     0.1353        408        640:  46%|████▌     | 369/809 [07:14<08:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.173     0.4403      0.135        395        640:  46%|████▌     | 371/809 [07:17<08:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.173     0.4403     0.1349        364        640:  46%|████▌     | 373/809 [07:19<08:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.172     0.4405     0.1346        356        640:  46%|████▋     | 375/809 [07:21<08:20,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.172     0.4403     0.1347        318        640:  46%|████▋     | 376/809 [07:23<08:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.17     0.4407     0.1344        175        640:  47%|████▋     | 378/809 [07:25<08:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.171     0.4403     0.1342        277        640:  47%|████▋     | 380/809 [07:27<08:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.172     0.4398     0.1344        293        640:  47%|████▋     | 381/809 [07:28<08:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.175      0.439      0.135        504        640:  47%|████▋     | 383/809 [07:31<08:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.176     0.4385     0.1349        251        640:  48%|████▊     | 385/809 [07:33<08:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.176     0.4387     0.1347         88        640:  48%|████▊     | 387/809 [07:36<08:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.175      0.439     0.1346        114        640:  48%|████▊     | 388/809 [07:37<08:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.176     0.4383     0.1347        267        640:  48%|████▊     | 390/809 [07:39<08:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4379     0.1353        427        640:  48%|████▊     | 392/809 [07:41<08:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4379     0.1352        171        640:  49%|████▊     | 393/809 [07:43<08:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.177     0.4382     0.1351        223        640:  49%|████▉     | 395/809 [07:45<08:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.177     0.4381      0.135        182        640:  49%|████▉     | 397/809 [07:47<08:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4379     0.1349        266        640:  49%|████▉     | 398/809 [07:49<08:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.177      0.438     0.1348        222        640:  49%|████▉     | 400/809 [07:51<07:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.177      0.438     0.1348        295        640:  50%|████▉     | 402/809 [07:53<07:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.176     0.4381     0.1347        124        640:  50%|████▉     | 404/809 [07:56<08:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.176     0.4381     0.1346        230        640:  50%|█████     | 405/809 [07:57<07:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.177     0.4377     0.1346        218        640:  50%|█████     | 407/809 [07:59<07:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4375     0.1347        345        640:  51%|█████     | 409/809 [08:01<07:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4373     0.1346        242        640:  51%|█████     | 410/809 [08:03<07:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4373     0.1346        190        640:  51%|█████     | 412/809 [08:05<07:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4368     0.1352        337        640:  51%|█████     | 414/809 [08:07<07:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4368     0.1351        209        640:  51%|█████▏    | 416/809 [08:10<07:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4372      0.135        117        640:  52%|█████▏    | 417/809 [08:11<07:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4369     0.1351        248        640:  52%|█████▏    | 419/809 [08:13<07:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4367     0.1349        264        640:  52%|█████▏    | 421/809 [08:15<07:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4366      0.135        400        640:  52%|█████▏    | 422/809 [08:17<07:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4367     0.1348        240        640:  52%|█████▏    | 424/809 [08:19<07:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4365     0.1348        283        640:  53%|█████▎    | 426/809 [08:21<07:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4366     0.1346        156        640:  53%|█████▎    | 428/809 [08:24<07:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4368     0.1345        231        640:  53%|█████▎    | 429/809 [08:25<07:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4363     0.1346        230        640:  53%|█████▎    | 431/809 [08:27<07:41,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4363     0.1346        418        640:  54%|█████▎    | 433/809 [08:30<07:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4363     0.1345        259        640:  54%|█████▎    | 434/809 [08:31<07:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4366     0.1343        116        640:  54%|█████▍    | 436/809 [08:33<07:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4371     0.1341        133        640:  54%|█████▍    | 438/809 [08:36<07:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4368     0.1342        504        640:  54%|█████▍    | 439/809 [08:37<07:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178      0.437     0.1341        246        640:  55%|█████▍    | 441/809 [08:39<07:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4368     0.1344        413        640:  55%|█████▍    | 443/809 [08:41<07:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4367     0.1344        336        640:  55%|█████▍    | 444/809 [08:43<07:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4366     0.1344        252        640:  55%|█████▌    | 446/809 [08:45<07:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4369     0.1342        345        640:  55%|█████▌    | 448/809 [08:47<07:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4371      0.134        198        640:  56%|█████▌    | 449/809 [08:49<06:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.177     0.4374     0.1338        308        640:  56%|█████▌    | 451/809 [08:51<06:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4373     0.1338        385        640:  56%|█████▌    | 453/809 [08:53<06:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.176     0.4378     0.1335        160        640:  56%|█████▌    | 455/809 [08:56<06:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.176     0.4379     0.1334        131        640:  56%|█████▋    | 456/809 [08:57<06:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.176      0.438     0.1333        194        640:  57%|█████▋    | 458/809 [08:59<06:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.174     0.4384     0.1331        159        640:  57%|█████▋    | 460/809 [09:01<06:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.174     0.4386      0.133        134        640:  57%|█████▋    | 461/809 [09:03<06:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.175     0.4383     0.1329        280        640:  57%|█████▋    | 463/809 [09:05<06:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.175     0.4382     0.1329        476        640:  57%|█████▋    | 465/809 [09:07<06:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.175     0.4382     0.1327        166        640:  58%|█████▊    | 467/809 [09:10<06:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.175     0.4381     0.1327        158        640:  58%|█████▊    | 468/809 [09:11<07:05,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.176     0.4375     0.1327        212        640:  58%|█████▊    | 470/809 [09:13<06:52,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.177      0.437     0.1327        139        640:  58%|█████▊    | 472/809 [09:16<06:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4368     0.1327        350        640:  58%|█████▊    | 473/809 [09:17<06:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4366     0.1325        247        640:  59%|█████▊    | 475/809 [09:19<06:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4365     0.1326        429        640:  59%|█████▉    | 477/809 [09:22<06:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4362     0.1328        309        640:  59%|█████▉    | 478/809 [09:23<06:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4361     0.1328        146        640:  59%|█████▉    | 480/809 [09:25<06:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4361     0.1326        221        640:  60%|█████▉    | 482/809 [09:28<06:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4361     0.1326        142        640:  60%|█████▉    | 483/809 [09:29<06:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4361     0.1325        212        640:  60%|█████▉    | 485/809 [09:31<06:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4361     0.1325        220        640:  60%|██████    | 487/809 [09:33<06:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4354     0.1328        372        640:  60%|██████    | 489/809 [09:36<06:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4352     0.1328        244        640:  61%|██████    | 490/809 [09:37<06:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4352      0.133        180        640:  61%|██████    | 492/809 [09:39<06:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4351      0.133        182        640:  61%|██████    | 494/809 [09:42<06:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4349     0.1331        193        640:  61%|██████    | 495/809 [09:43<06:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4349     0.1331        236        640:  61%|██████▏   | 497/809 [09:45<06:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4349      0.133        324        640:  62%|██████▏   | 499/809 [09:48<06:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4347     0.1331        491        640:  62%|██████▏   | 500/809 [09:49<06:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.183     0.4344     0.1333        281        640:  62%|██████▏   | 502/809 [09:51<06:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.183     0.4344     0.1332        365        640:  62%|██████▏   | 504/809 [09:54<05:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.183     0.4342     0.1332        386        640:  62%|██████▏   | 505/809 [09:55<05:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.183     0.4344     0.1331        174        640:  63%|██████▎   | 507/809 [09:57<05:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.183     0.4342     0.1331        378        640:  63%|██████▎   | 509/809 [09:59<05:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4338     0.1331        385        640:  63%|██████▎   | 511/809 [10:02<05:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4337     0.1331        225        640:  63%|██████▎   | 512/809 [10:03<05:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4338      0.133        241        640:  64%|██████▎   | 514/809 [10:05<05:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4335     0.1332        221        640:  64%|██████▍   | 516/809 [10:08<05:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4335     0.1331        316        640:  64%|██████▍   | 518/809 [10:10<05:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4333     0.1333        330        640:  64%|██████▍   | 519/809 [10:11<05:34,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4332     0.1332        236        640:  64%|██████▍   | 521/809 [10:13<05:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4331     0.1332        247        640:  65%|██████▍   | 523/809 [10:16<05:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4331     0.1331        230        640:  65%|██████▍   | 525/809 [10:18<05:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.186     0.4329     0.1332        296        640:  65%|██████▌   | 526/809 [10:19<05:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4327     0.1336        179        640:  65%|██████▌   | 528/809 [10:22<05:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4321     0.1338        281        640:  66%|██████▌   | 530/809 [10:24<05:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4322     0.1336        170        640:  66%|██████▌   | 532/809 [10:26<05:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4322     0.1336        237        640:  66%|██████▌   | 533/809 [10:28<05:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.186     0.4322     0.1336        288        640:  66%|██████▌   | 535/809 [10:30<05:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.186     0.4322     0.1336        436        640:  66%|██████▋   | 537/809 [10:32<05:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4323     0.1335        132        640:  67%|██████▋   | 538/809 [10:34<05:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4324     0.1334        365        640:  67%|██████▋   | 540/809 [10:36<05:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4324     0.1334        478        640:  67%|██████▋   | 542/809 [10:38<05:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4323     0.1334        392        640:  67%|██████▋   | 543/809 [10:39<05:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4322     0.1335        168        640:  67%|██████▋   | 545/809 [10:42<05:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4321     0.1336        350        640:  68%|██████▊   | 547/809 [10:44<05:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4321     0.1337        191        640:  68%|██████▊   | 549/809 [10:47<05:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4323     0.1335        200        640:  68%|██████▊   | 550/809 [10:48<05:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4323     0.1335        251        640:  68%|██████▊   | 552/809 [10:50<05:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185      0.432     0.1336        424        640:  68%|██████▊   | 554/809 [10:53<05:10,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185      0.432     0.1335        231        640:  69%|██████▊   | 555/809 [10:54<05:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4321     0.1336        295        640:  69%|██████▉   | 557/809 [10:56<05:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4317     0.1337        299        640:  69%|██████▉   | 559/809 [10:59<04:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4316     0.1337        165        640:  69%|██████▉   | 560/809 [11:00<05:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.186     0.4319     0.1335        142        640:  69%|██████▉   | 562/809 [11:02<04:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.186      0.432     0.1334        328        640:  70%|██████▉   | 564/809 [11:04<04:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.186      0.432     0.1334        330        640:  70%|██████▉   | 565/809 [11:06<04:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4324     0.1332        170        640:  70%|███████   | 567/809 [11:08<04:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4324     0.1334        423        640:  70%|███████   | 569/809 [11:10<04:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4325     0.1333        153        640:  70%|███████   | 570/809 [11:12<04:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4324     0.1333        139        640:  71%|███████   | 572/809 [11:14<04:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4322     0.1334        155        640:  71%|███████   | 574/809 [11:16<04:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4323     0.1334        147        640:  71%|███████   | 576/809 [11:19<04:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4325     0.1334        145        640:  71%|███████▏  | 577/809 [11:20<04:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4324     0.1333        190        640:  72%|███████▏  | 579/809 [11:22<04:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4326     0.1332        119        640:  72%|███████▏  | 581/809 [11:25<04:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4328     0.1331         80        640:  72%|███████▏  | 582/809 [11:26<04:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4329      0.133        214        640:  72%|███████▏  | 584/809 [11:28<04:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.183     0.4331     0.1329        247        640:  72%|███████▏  | 586/809 [11:30<04:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4332     0.1328        226        640:  73%|███████▎  | 587/809 [11:32<04:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182     0.4333     0.1327        244        640:  73%|███████▎  | 589/809 [11:34<04:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4337     0.1326        155        640:  73%|███████▎  | 591/809 [11:36<04:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.181     0.4339     0.1325        318        640:  73%|███████▎  | 593/809 [11:39<04:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18      0.434     0.1324        156        640:  73%|███████▎  | 594/809 [11:40<04:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18      0.434     0.1324        140        640:  74%|███████▎  | 596/809 [11:42<04:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4341     0.1324        263        640:  74%|███████▍  | 597/809 [11:44<04:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.179     0.4344     0.1323         63        640:  74%|███████▍  | 599/809 [11:46<04:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4346     0.1322        268        640:  74%|███████▍  | 601/809 [11:48<04:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4346     0.1321        320        640:  75%|███████▍  | 603/809 [11:51<04:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4346      0.132        137        640:  75%|███████▍  | 604/809 [11:52<04:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4347     0.1319        334        640:  75%|███████▍  | 606/809 [11:54<03:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.177     0.4347     0.1318        249        640:  75%|███████▌  | 608/809 [11:57<03:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4345     0.1319        267        640:  75%|███████▌  | 609/809 [11:58<03:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.178     0.4343      0.132        233        640:  76%|███████▌  | 611/809 [12:00<03:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4338     0.1321        358        640:  76%|███████▌  | 613/809 [12:03<03:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4337     0.1321        326        640:  76%|███████▌  | 614/809 [12:04<03:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.18     0.4335     0.1322        266        640:  76%|███████▌  | 616/809 [12:06<03:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182      0.433     0.1325        373        640:  76%|███████▋  | 618/809 [12:08<03:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.182      0.433     0.1324        149        640:  77%|███████▋  | 620/809 [12:11<03:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.183     0.4327     0.1328        308        640:  77%|███████▋  | 621/809 [12:12<03:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.183     0.4325     0.1331        247        640:  77%|███████▋  | 623/809 [12:14<03:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4323     0.1332        335        640:  77%|███████▋  | 625/809 [12:17<03:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.185     0.4321     0.1334        401        640:  77%|███████▋  | 626/809 [12:18<03:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.184     0.4324     0.1335         82        640:  78%|███████▊  | 628/809 [12:20<03:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.186      0.432     0.1339        256        640:  78%|███████▊  | 630/809 [12:23<03:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.186     0.4319      0.134        301        640:  78%|███████▊  | 631/809 [12:24<03:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4319     0.1341        210        640:  78%|███████▊  | 633/809 [12:26<03:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4318     0.1341        236        640:  78%|███████▊  | 635/809 [12:29<03:37,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4317     0.1341        346        640:  79%|███████▊  | 636/809 [12:30<03:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4319      0.134         95        640:  79%|███████▉  | 638/809 [12:32<03:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4319      0.134        276        640:  79%|███████▉  | 640/809 [12:35<03:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4318      0.134        342        640:  79%|███████▉  | 641/809 [12:36<03:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4316      0.134        259        640:  79%|███████▉  | 643/809 [12:38<03:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4314     0.1339        308        640:  80%|███████▉  | 645/809 [12:41<03:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4312     0.1339        372        640:  80%|███████▉  | 647/809 [12:43<03:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4311      0.134        280        640:  80%|████████  | 648/809 [12:44<03:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4311     0.1339        185        640:  80%|████████  | 650/809 [12:46<03:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4311     0.1338        110        640:  81%|████████  | 652/809 [12:49<03:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4311     0.1338        205        640:  81%|████████  | 653/809 [12:50<03:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4309     0.1338        478        640:  81%|████████  | 655/809 [12:52<02:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4309     0.1338        306        640:  81%|████████  | 657/809 [12:55<02:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.187     0.4309     0.1338        127        640:  81%|████████▏ | 658/809 [12:56<03:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.188     0.4307     0.1338        218        640:  82%|████████▏ | 660/809 [12:58<02:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.189     0.4307     0.1339        142        640:  82%|████████▏ | 662/809 [13:01<02:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.189     0.4308     0.1339        179        640:  82%|████████▏ | 663/809 [13:02<03:00,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.189      0.431     0.1339        234        640:  82%|████████▏ | 665/809 [13:04<02:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.19     0.4307     0.1341        260        640:  82%|████████▏ | 667/809 [13:07<02:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G       1.19     0.4306     0.1342        304        640:  83%|████████▎ | 669/809 [13:09<02:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.191     0.4305     0.1344        369        640:  83%|████████▎ | 670/809 [13:10<02:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.191     0.4306     0.1344        104        640:  83%|████████▎ | 672/809 [13:12<02:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.192     0.4303     0.1353        232        640:  83%|████████▎ | 674/809 [13:15<02:44,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4304     0.1354        260        640:  84%|████████▎ | 676/809 [13:18<02:47,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4304     0.1354        182        640:  84%|████████▎ | 677/809 [13:19<02:43,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4305     0.1354        202        640:  84%|████████▍ | 679/809 [13:21<02:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4305     0.1355        202        640:  84%|████████▍ | 680/809 [13:22<02:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4304     0.1358        117        640:  84%|████████▍ | 682/809 [13:25<02:38,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4304     0.1358        166        640:  85%|████████▍ | 684/809 [13:27<02:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4303     0.1358        121        640:  85%|████████▍ | 686/809 [13:29<02:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4303     0.1358        291        640:  85%|████████▍ | 687/809 [13:31<02:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4302     0.1359        240        640:  85%|████████▌ | 689/809 [13:33<02:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195       0.43      0.136        236        640:  85%|████████▌ | 691/809 [13:35<02:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4301     0.1359        142        640:  86%|████████▌ | 692/809 [13:37<02:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4303     0.1359        354        640:  86%|████████▌ | 694/809 [13:39<02:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4302      0.136        136        640:  86%|████████▌ | 696/809 [13:41<02:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4299     0.1361        345        640:  86%|████████▋ | 698/809 [13:43<02:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4299     0.1362        237        640:  86%|████████▋ | 699/809 [13:45<02:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4299     0.1362        213        640:  87%|████████▋ | 701/809 [13:47<02:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196       0.43     0.1362        240        640:  87%|████████▋ | 703/809 [13:49<02:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4302     0.1361         81        640:  87%|████████▋ | 705/809 [13:52<01:59,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4301     0.1362        365        640:  87%|████████▋ | 706/809 [13:53<01:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4305      0.136        195        640:  88%|████████▊ | 708/809 [13:55<01:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4306      0.136        152        640:  88%|████████▊ | 710/809 [13:57<01:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4306      0.136        269        640:  88%|████████▊ | 712/809 [14:00<01:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4306     0.1361        225        640:  88%|████████▊ | 713/809 [14:01<01:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4308      0.136        264        640:  88%|████████▊ | 715/809 [14:03<01:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4309     0.1359        157        640:  89%|████████▊ | 716/809 [14:05<01:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4308     0.1364        274        640:  89%|████████▉ | 718/809 [14:07<01:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195      0.431     0.1363        170        640:  89%|████████▉ | 720/809 [14:09<01:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4309     0.1363        310        640:  89%|████████▉ | 722/809 [14:12<01:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4309     0.1363        240        640:  89%|████████▉ | 723/809 [14:13<01:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4308     0.1363        132        640:  90%|████████▉ | 725/809 [14:15<01:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4307     0.1363        120        640:  90%|████████▉ | 727/809 [14:17<01:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.197     0.4304     0.1365        316        640:  90%|████████▉ | 728/809 [14:19<01:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.197     0.4303     0.1365        111        640:  90%|█████████ | 730/809 [14:21<01:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.197     0.4305     0.1364        234        640:  90%|█████████ | 732/809 [14:24<01:33,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4306     0.1363        211        640:  91%|█████████ | 733/809 [14:25<01:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4306     0.1363        429        640:  91%|█████████ | 735/809 [14:27<01:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.196     0.4306     0.1362        186        640:  91%|█████████ | 737/809 [14:29<01:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4307     0.1361        207        640:  91%|█████████ | 738/809 [14:31<01:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4308     0.1361        180        640:  91%|█████████▏| 740/809 [14:33<01:22,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4309     0.1362        392        640:  92%|█████████▏| 742/809 [14:35<01:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4307     0.1362        266        640:  92%|█████████▏| 743/809 [14:37<01:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4309     0.1361        206        640:  92%|█████████▏| 745/809 [14:39<01:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4309     0.1361        302        640:  92%|█████████▏| 747/809 [14:41<01:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4312     0.1359        305        640:  93%|█████████▎| 749/809 [14:44<01:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4313     0.1359        230        640:  93%|█████████▎| 750/809 [14:45<01:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.192     0.4315     0.1358        344        640:  93%|█████████▎| 752/809 [14:47<01:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.192     0.4314     0.1358        158        640:  93%|█████████▎| 754/809 [14:50<01:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4312     0.1359        354        640:  93%|█████████▎| 755/809 [14:51<01:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.192     0.4314     0.1358        198        640:  94%|█████████▎| 757/809 [14:53<01:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4313     0.1358        426        640:  94%|█████████▍| 759/809 [14:55<00:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4313     0.1358        181        640:  94%|█████████▍| 760/809 [14:57<00:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4314     0.1359        128        640:  94%|█████████▍| 762/809 [14:59<00:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.192     0.4314     0.1358        245        640:  94%|█████████▍| 764/809 [15:01<00:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.192     0.4314     0.1358        373        640:  95%|█████████▍| 765/809 [15:03<00:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4311     0.1364        269        640:  95%|█████████▍| 767/809 [15:05<00:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4312     0.1364        160        640:  95%|█████████▌| 769/809 [15:07<00:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4312     0.1365        434        640:  95%|█████████▌| 771/809 [15:10<00:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4313     0.1366        142        640:  95%|█████████▌| 772/809 [15:11<00:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4312     0.1367        279        640:  96%|█████████▌| 774/809 [15:13<00:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4311     0.1366        177        640:  96%|█████████▌| 776/809 [15:16<00:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4311     0.1366        159        640:  96%|█████████▌| 777/809 [15:17<00:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4312     0.1365        133        640:  96%|█████████▋| 779/809 [15:19<00:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4315     0.1365         49        640:  97%|█████████▋| 781/809 [15:22<00:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4313     0.1366        426        640:  97%|█████████▋| 782/809 [15:23<00:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4312     0.1366        216        640:  97%|█████████▋| 784/809 [15:25<00:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4312     0.1365        181        640:  97%|█████████▋| 786/809 [15:28<00:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.195     0.4312     0.1365        142        640:  97%|█████████▋| 788/809 [15:30<00:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4312     0.1364        253        640:  98%|█████████▊| 789/809 [15:31<00:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4315     0.1364        395        640:  98%|█████████▊| 791/809 [15:33<00:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.194     0.4315     0.1363        128        640:  98%|█████████▊| 793/809 [15:36<00:19,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4317     0.1363         78        640:  98%|█████████▊| 794/809 [15:37<00:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.193     0.4319     0.1362        233        640:  98%|█████████▊| 796/809 [15:39<00:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.192     0.4322     0.1361        161        640:  99%|█████████▊| 798/809 [15:42<00:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.192     0.4322      0.136        202        640:  99%|█████████▉| 799/809 [15:43<00:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.191     0.4323      0.136        361        640:  99%|█████████▉| 801/809 [15:45<00:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.191     0.4323      0.136        153        640:  99%|█████████▉| 803/809 [15:48<00:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.191     0.4322     0.1359        207        640:  99%|█████████▉| 804/809 [15:49<00:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.192     0.4321      0.136        463        640: 100%|█████████▉| 806/809 [15:51<00:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     84/300      14.3G      1.191     0.4324      0.136         72        640: 100%|█████████▉| 808/809 [15:53<00:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   7%|▋         | 5/69 [00:01<00:15,  4.17it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  19%|█▉        | 13/69 [00:03<00:13,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  30%|███       | 21/69 [00:05<00:11,  4.06it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  42%|████▏     | 29/69 [00:07<00:10,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  54%|█████▎    | 37/69 [00:09<00:08,  3.97it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  65%|██████▌   | 45/69 [00:11<00:06,  3.81it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  75%|███████▌  | 52/69 [00:13<00:04,  3.96it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  87%|████████▋ | 60/69 [00:15<00:02,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                   all        548      38759      0.529      0.353      0.357      0.193
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.249      0.418     0.1234        318        640:   0%|          | 2/809 [00:02<15:41,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.157     0.4458     0.1265        441        640:   0%|          | 4/809 [00:04<15:32,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.146     0.4461     0.1243        382        640:   1%|          | 5/809 [00:05<15:30,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.159      0.456     0.1226        395        640:   1%|          | 7/809 [00:08<15:53,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.124     0.4628     0.1137        249        640:   1%|          | 9/809 [00:10<16:00,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.093     0.4686     0.1161         95        640:   1%|▏         | 11/809 [00:12<15:41,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.12     0.4579     0.1385         78        640:   2%|▏         | 13/809 [00:15<15:40,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.115      0.472     0.1409         56        640:   2%|▏         | 14/809 [00:16<15:47,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.132      0.467     0.1477        545        640:   2%|▏         | 16/809 [00:18<15:42,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.107     0.4722     0.1396        311        640:   2%|▏         | 18/809 [00:21<15:36,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.11     0.4713     0.1381        292        640:   2%|▏         | 19/809 [00:22<16:29,  1.25s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.099     0.4725     0.1325        175        640:   3%|▎         | 21/809 [00:25<15:59,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.121     0.4653     0.1392        530        640:   3%|▎         | 22/809 [00:26<15:56,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.105     0.4704     0.1364        114        640:   3%|▎         | 24/809 [00:28<15:52,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.108     0.4704     0.1348        197        640:   3%|▎         | 26/809 [00:31<15:46,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.119     0.4674      0.135        416        640:   3%|▎         | 27/809 [00:32<15:53,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.129     0.4646     0.1358        185        640:   4%|▎         | 29/809 [00:34<15:24,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.138     0.4635     0.1354        284        640:   4%|▍         | 31/809 [00:37<15:30,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.126     0.4677     0.1334        155        640:   4%|▍         | 32/809 [00:38<15:17,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.129      0.467     0.1343        458        640:   4%|▍         | 34/809 [00:40<15:17,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.116     0.4727     0.1319        215        640:   4%|▍         | 36/809 [00:42<15:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.105     0.4771     0.1296        254        640:   5%|▍         | 38/809 [00:45<14:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.105     0.4765     0.1298        460        640:   5%|▍         | 39/809 [00:46<14:57,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.104     0.4791       0.13        273        640:   5%|▌         | 41/809 [00:48<15:13,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.113     0.4753     0.1305        132        640:   5%|▌         | 43/809 [00:51<14:56,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.125     0.4712     0.1316        621        640:   5%|▌         | 44/809 [00:52<15:00,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.134     0.4674     0.1312        159        640:   6%|▌         | 46/809 [00:54<14:45,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.133     0.4669     0.1301        142        640:   6%|▌         | 48/809 [00:57<14:59,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.135     0.4656     0.1297        392        640:   6%|▌         | 50/809 [00:59<14:58,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.132     0.4659     0.1288        311        640:   6%|▋         | 51/809 [01:00<14:53,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.126     0.4673     0.1269        238        640:   7%|▋         | 53/809 [01:03<15:05,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.138      0.463     0.1314        397        640:   7%|▋         | 55/809 [01:05<15:00,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.14     0.4623     0.1316        447        640:   7%|▋         | 56/809 [01:06<14:50,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.132     0.4647     0.1305        335        640:   7%|▋         | 58/809 [01:08<14:39,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.128     0.4657     0.1288        218        640:   7%|▋         | 60/809 [01:11<14:28,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.126     0.4669     0.1288        394        640:   8%|▊         | 62/809 [01:13<14:20,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.125      0.467     0.1284        194        640:   8%|▊         | 63/809 [01:14<14:46,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.123      0.468      0.129        391        640:   8%|▊         | 65/809 [01:17<14:52,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.128      0.466     0.1286        213        640:   8%|▊         | 66/809 [01:18<15:19,  1.24s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.127     0.4661     0.1273        182        640:   8%|▊         | 68/809 [01:20<14:57,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.129     0.4653      0.127        404        640:   9%|▊         | 70/809 [01:23<14:43,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.12     0.4683     0.1259        159        640:   9%|▉         | 72/809 [01:25<14:18,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.124      0.467     0.1256        271        640:   9%|▉         | 73/809 [01:26<14:19,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.123     0.4669     0.1252        225        640:   9%|▉         | 75/809 [01:29<14:33,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.125     0.4666     0.1256        270        640:  10%|▉         | 77/809 [01:31<14:20,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.132     0.4646     0.1268        645        640:  10%|▉         | 78/809 [01:32<14:39,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.128     0.4649     0.1257         97        640:  10%|▉         | 80/809 [01:35<14:23,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.13     0.4643      0.126        149        640:  10%|█         | 82/809 [01:37<14:17,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.132     0.4632     0.1257        298        640:  10%|█         | 83/809 [01:38<14:08,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.134     0.4633     0.1263        216        640:  11%|█         | 85/809 [01:40<13:56,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.132     0.4634     0.1253        333        640:  11%|█         | 87/809 [01:43<13:51,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.132     0.4646     0.1259        229        640:  11%|█         | 89/809 [01:45<14:11,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.13     0.4653     0.1254        175        640:  11%|█         | 90/809 [01:46<14:26,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.125     0.4664     0.1247        214        640:  11%|█▏        | 92/809 [01:49<14:11,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.12     0.4681     0.1238         90        640:  12%|█▏        | 94/809 [01:51<14:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.12     0.4682     0.1237        171        640:  12%|█▏        | 95/809 [01:52<13:56,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.115     0.4694      0.123        285        640:  12%|█▏        | 97/809 [01:54<14:00,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.114     0.4696     0.1228        309        640:  12%|█▏        | 99/809 [01:57<13:59,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.111     0.4703      0.122        139        640:  12%|█▏        | 101/809 [01:59<13:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.107     0.4711     0.1215         81        640:  13%|█▎        | 102/809 [02:00<13:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.106     0.4718     0.1211        224        640:  13%|█▎        | 104/809 [02:03<13:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.108     0.4712     0.1217        370        640:  13%|█▎        | 106/809 [02:05<13:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.111     0.4704      0.122        139        640:  13%|█▎        | 107/809 [02:06<13:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.114     0.4695     0.1227        168        640:  13%|█▎        | 109/809 [02:08<13:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.119      0.468     0.1244        260        640:  14%|█▎        | 111/809 [02:11<13:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.119     0.4677     0.1242        291        640:  14%|█▍        | 112/809 [02:12<13:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.12     0.4671     0.1248        214        640:  14%|█▍        | 114/809 [02:14<13:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.118     0.4677     0.1249        164        640:  14%|█▍        | 116/809 [02:17<13:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.118     0.4679     0.1247        359        640:  15%|█▍        | 118/809 [02:19<13:17,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.117     0.4677     0.1245        300        640:  15%|█▍        | 119/809 [02:20<13:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.113      0.468     0.1244        145        640:  15%|█▍        | 121/809 [02:23<14:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.111     0.4679     0.1238        195        640:  15%|█▌        | 123/809 [02:25<13:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.112     0.4675     0.1238        246        640:  15%|█▌        | 124/809 [02:26<13:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.11     0.4681     0.1236        164        640:  16%|█▌        | 126/809 [02:29<13:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.104     0.4702      0.123         67        640:  16%|█▌        | 128/809 [02:31<13:41,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.102     0.4706     0.1229        200        640:  16%|█▌        | 129/809 [02:32<13:48,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G        1.1     0.4707     0.1224        306        640:  16%|█▌        | 131/809 [02:35<13:46,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.095     0.4716     0.1219        188        640:  16%|█▋        | 133/809 [02:37<13:45,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.095     0.4717     0.1217        306        640:  17%|█▋        | 134/809 [02:38<13:42,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.095     0.4715     0.1221        698        640:  17%|█▋        | 136/809 [02:41<14:05,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.093     0.4719     0.1219        192        640:  17%|█▋        | 137/809 [02:42<13:44,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.093     0.4714     0.1232        105        640:  17%|█▋        | 139/809 [02:45<13:40,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.094     0.4711     0.1233        249        640:  17%|█▋        | 141/809 [02:47<13:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.094     0.4707     0.1233        526        640:  18%|█▊        | 143/809 [02:49<13:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.097     0.4697     0.1245        370        640:  18%|█▊        | 144/809 [02:50<13:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.094       0.47      0.124        279        640:  18%|█▊        | 146/809 [02:53<12:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4719     0.1233        140        640:  18%|█▊        | 148/809 [02:55<12:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087      0.472      0.123        350        640:  18%|█▊        | 149/809 [02:56<12:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4721     0.1227        385        640:  19%|█▊        | 151/809 [02:58<12:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4716     0.1223        366        640:  19%|█▉        | 153/809 [03:01<12:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4723     0.1218        139        640:  19%|█▉        | 155/809 [03:03<12:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081      0.472     0.1217        239        640:  19%|█▉        | 156/809 [03:04<12:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4721     0.1213        242        640:  20%|█▉        | 158/809 [03:07<12:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4718     0.1209        270        640:  20%|█▉        | 160/809 [03:09<13:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4715     0.1212        364        640:  20%|█▉        | 161/809 [03:10<13:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4719      0.121        484        640:  20%|██        | 163/809 [03:13<12:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4721     0.1207        302        640:  20%|██        | 165/809 [03:15<12:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4708     0.1209        388        640:  21%|██        | 167/809 [03:17<12:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4701      0.122        323        640:  21%|██        | 169/809 [03:20<12:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4697      0.122        365        640:  21%|██        | 170/809 [03:21<12:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4694     0.1217        247        640:  21%|██▏       | 172/809 [03:23<12:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4698     0.1214        123        640:  22%|██▏       | 174/809 [03:26<12:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4689     0.1216        180        640:  22%|██▏       | 176/809 [03:28<12:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.088     0.4684     0.1216        419        640:  22%|██▏       | 177/809 [03:29<12:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.088      0.468     0.1213        157        640:  22%|██▏       | 179/809 [03:31<12:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.089      0.468      0.121        137        640:  22%|██▏       | 181/809 [03:34<12:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.089     0.4674     0.1207        335        640:  23%|██▎       | 183/809 [03:36<11:55,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.09      0.467     0.1207        177        640:  23%|██▎       | 184/809 [03:37<11:56,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.088     0.4674     0.1204        169        640:  23%|██▎       | 186/809 [03:39<12:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4681       0.12        246        640:  23%|██▎       | 188/809 [03:42<12:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4682     0.1199        159        640:  23%|██▎       | 189/809 [03:43<11:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4689     0.1193        173        640:  24%|██▎       | 191/809 [03:45<11:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4694     0.1191        155        640:  24%|██▍       | 193/809 [03:48<12:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4699      0.119         95        640:  24%|██▍       | 194/809 [03:49<12:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4702     0.1186         91        640:  24%|██▍       | 196/809 [03:51<12:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4709     0.1183        146        640:  24%|██▍       | 198/809 [03:54<11:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4706     0.1185        297        640:  25%|██▍       | 200/809 [03:56<11:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4708     0.1184         82        640:  25%|██▍       | 201/809 [03:57<12:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4707     0.1185        358        640:  25%|██▌       | 203/809 [03:59<11:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4709     0.1183        226        640:  25%|██▌       | 204/809 [04:02<11:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081      0.471     0.1181        132        640:  25%|██▌       | 206/809 [04:03<11:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081      0.471      0.118        192        640:  26%|██▌       | 208/809 [04:05<12:00,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4708      0.118        314        640:  26%|██▌       | 210/809 [04:08<11:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4711      0.118        233        640:  26%|██▌       | 211/809 [04:09<11:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4716     0.1178        229        640:  26%|██▋       | 213/809 [04:11<11:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4718     0.1181        464        640:  27%|██▋       | 215/809 [04:14<11:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4721      0.118        192        640:  27%|██▋       | 216/809 [04:15<11:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4726     0.1177        181        640:  27%|██▋       | 218/809 [04:17<11:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4735     0.1175         88        640:  27%|██▋       | 220/809 [04:20<11:17,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4735     0.1174        255        640:  27%|██▋       | 222/809 [04:22<11:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.077     0.4737     0.1172        187        640:  28%|██▊       | 223/809 [04:23<11:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.077     0.4735      0.117        314        640:  28%|██▊       | 225/809 [04:25<11:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.077     0.4732     0.1169        227        640:  28%|██▊       | 227/809 [04:28<11:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.076     0.4735     0.1168         83        640:  28%|██▊       | 229/809 [04:30<11:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.075     0.4738     0.1168        249        640:  28%|██▊       | 230/809 [04:31<11:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078      0.473     0.1169        368        640:  29%|██▊       | 232/809 [04:33<11:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4724      0.117        336        640:  29%|██▉       | 234/809 [04:36<11:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4724     0.1169        424        640:  29%|██▉       | 236/809 [04:38<11:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4727     0.1168        200        640:  29%|██▉       | 237/809 [04:39<10:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4724     0.1167        174        640:  30%|██▉       | 239/809 [04:42<11:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4719     0.1168        469        640:  30%|██▉       | 241/809 [04:44<11:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4718     0.1168        124        640:  30%|██▉       | 242/809 [04:45<11:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078      0.472     0.1165        232        640:  30%|███       | 244/809 [04:47<10:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.076     0.4724     0.1164        150        640:  30%|███       | 246/809 [04:50<10:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.074     0.4729     0.1162         92        640:  31%|███       | 247/809 [04:51<11:12,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.072     0.4733     0.1158        211        640:  31%|███       | 249/809 [04:53<11:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.073     0.4732     0.1158        267        640:  31%|███       | 251/809 [04:56<10:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.073     0.4732     0.1156        252        640:  31%|███▏      | 253/809 [04:58<10:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.072     0.4735     0.1158        136        640:  31%|███▏      | 254/809 [04:59<10:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.071      0.474     0.1156        147        640:  32%|███▏      | 256/809 [05:01<10:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.072     0.4736     0.1158        174        640:  32%|███▏      | 258/809 [05:04<10:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.071      0.474     0.1157        122        640:  32%|███▏      | 259/809 [05:05<11:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.073     0.4734     0.1158        285        640:  32%|███▏      | 261/809 [05:08<10:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.07     0.4742     0.1156         50        640:  33%|███▎      | 263/809 [05:10<10:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.07     0.4741     0.1155        314        640:  33%|███▎      | 264/809 [05:11<10:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.069     0.4742     0.1153        111        640:  33%|███▎      | 266/809 [05:13<10:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.07     0.4739     0.1152        134        640:  33%|███▎      | 268/809 [05:16<10:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.069     0.4739      0.115        178        640:  33%|███▎      | 269/809 [05:18<10:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.07     0.4739      0.115        191        640:  33%|███▎      | 271/809 [05:19<10:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.072      0.473     0.1154        194        640:  34%|███▎      | 273/809 [05:22<10:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.072     0.4728     0.1151        239        640:  34%|███▍      | 275/809 [05:24<10:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.071     0.4733      0.115         66        640:  34%|███▍      | 276/809 [05:25<10:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.071     0.4733     0.1149        125        640:  34%|███▍      | 278/809 [05:28<10:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.071     0.4731      0.115        509        640:  35%|███▍      | 280/809 [05:30<10:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.072     0.4727     0.1153        275        640:  35%|███▍      | 281/809 [05:31<10:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.072     0.4726     0.1151        173        640:  35%|███▍      | 283/809 [05:33<10:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.074     0.4719     0.1151        314        640:  35%|███▌      | 285/809 [05:36<10:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.073      0.472     0.1149        236        640:  35%|███▌      | 287/809 [05:38<10:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.075     0.4714     0.1149        563        640:  36%|███▌      | 288/809 [05:39<10:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.075     0.4714     0.1149        339        640:  36%|███▌      | 290/809 [05:42<10:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.077     0.4704     0.1156        219        640:  36%|███▌      | 292/809 [05:44<10:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4696     0.1163        438        640:  36%|███▋      | 294/809 [05:46<09:53,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4695     0.1162        179        640:  36%|███▋      | 295/809 [05:47<10:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4693     0.1168        395        640:  37%|███▋      | 297/809 [05:50<10:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4692     0.1167        268        640:  37%|███▋      | 299/809 [05:52<09:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4693     0.1166        333        640:  37%|███▋      | 300/809 [05:53<09:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4691      0.117        153        640:  37%|███▋      | 302/809 [05:55<09:45,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4694     0.1169        182        640:  38%|███▊      | 304/809 [05:58<09:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4696     0.1168        200        640:  38%|███▊      | 306/809 [06:00<09:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4698     0.1168        142        640:  38%|███▊      | 307/809 [06:01<09:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4704     0.1165         77        640:  38%|███▊      | 309/809 [06:04<09:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082       0.47     0.1169        643        640:  38%|███▊      | 311/809 [06:06<09:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082       0.47     0.1167        243        640:  39%|███▊      | 313/809 [06:09<09:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4696     0.1168        329        640:  39%|███▉      | 314/809 [06:10<09:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4697     0.1168        200        640:  39%|███▉      | 316/809 [06:12<09:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4701     0.1166        287        640:  39%|███▉      | 318/809 [06:15<09:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4699     0.1166        366        640:  39%|███▉      | 319/809 [06:16<09:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4705     0.1163         98        640:  40%|███▉      | 321/809 [06:18<09:46,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4709     0.1161        156        640:  40%|███▉      | 323/809 [06:20<09:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4712     0.1159        140        640:  40%|████      | 325/809 [06:23<09:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078      0.471     0.1159        375        640:  40%|████      | 326/809 [06:24<09:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4711     0.1157        384        640:  41%|████      | 328/809 [06:26<09:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4712     0.1156        144        640:  41%|████      | 330/809 [06:29<09:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4712     0.1155        188        640:  41%|████      | 331/809 [06:30<09:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082       0.47     0.1161        483        640:  41%|████      | 333/809 [06:32<09:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4698     0.1159        256        640:  41%|████▏     | 335/809 [06:35<09:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4698     0.1158        232        640:  42%|████▏     | 336/809 [06:36<09:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4696     0.1161        177        640:  42%|████▏     | 338/809 [06:38<09:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4692     0.1161        365        640:  42%|████▏     | 340/809 [06:41<09:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4694     0.1159        232        640:  42%|████▏     | 341/809 [06:42<09:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4695     0.1158        170        640:  42%|████▏     | 343/809 [06:44<09:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4692      0.116        230        640:  43%|████▎     | 345/809 [06:47<09:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4692     0.1159        295        640:  43%|████▎     | 347/809 [06:49<08:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4689      0.116        437        640:  43%|████▎     | 348/809 [06:50<09:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4691     0.1159        245        640:  43%|████▎     | 350/809 [06:52<09:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4693     0.1158        175        640:  44%|████▎     | 352/809 [06:55<08:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4693     0.1158        195        640:  44%|████▎     | 353/809 [06:56<08:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4698     0.1156        178        640:  44%|████▍     | 355/809 [06:58<08:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4697     0.1155        221        640:  44%|████▍     | 357/809 [07:01<08:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4698     0.1154        231        640:  44%|████▍     | 359/809 [07:03<08:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4693     0.1168        496        640:  44%|████▍     | 360/809 [07:04<08:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4693     0.1167        259        640:  45%|████▍     | 362/809 [07:07<08:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4687     0.1167        159        640:  45%|████▍     | 364/809 [07:09<08:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.088     0.4683     0.1166        118        640:  45%|████▌     | 365/809 [07:10<08:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.09     0.4676     0.1168        352        640:  45%|████▌     | 367/809 [07:12<08:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.093     0.4668     0.1171        211        640:  46%|████▌     | 369/809 [07:15<08:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.092     0.4669      0.117        252        640:  46%|████▌     | 370/809 [07:16<08:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.091      0.467     0.1168        270        640:  46%|████▌     | 372/809 [07:18<08:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.092     0.4671      0.117        301        640:  46%|████▌     | 374/809 [07:21<08:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.091     0.4673     0.1169        143        640:  46%|████▋     | 376/809 [07:23<08:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.091     0.4673      0.117        466        640:  47%|████▋     | 377/809 [07:24<08:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.09     0.4676     0.1168        108        640:  47%|████▋     | 379/809 [07:27<08:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.09     0.4678     0.1168        146        640:  47%|████▋     | 381/809 [07:29<08:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.089     0.4679     0.1167        187        640:  47%|████▋     | 382/809 [07:30<08:12,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.089     0.4681     0.1165        283        640:  47%|████▋     | 384/809 [07:32<08:06,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.088     0.4686     0.1164         46        640:  48%|████▊     | 386/809 [07:35<08:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4687     0.1163        255        640:  48%|████▊     | 387/809 [07:36<08:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.088     0.4685     0.1166        163        640:  48%|████▊     | 389/809 [07:38<08:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.088     0.4685     0.1166        135        640:  48%|████▊     | 391/809 [07:41<08:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.089      0.468     0.1166        360        640:  49%|████▊     | 393/809 [07:43<08:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.088     0.4681     0.1165        277        640:  49%|████▊     | 394/809 [07:44<08:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.089      0.468     0.1167        326        640:  49%|████▉     | 396/809 [07:46<08:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.088     0.4683     0.1166        302        640:  49%|████▉     | 398/809 [07:49<07:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4686     0.1164        253        640:  49%|████▉     | 400/809 [07:51<07:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4687     0.1164        152        640:  50%|████▉     | 401/809 [07:52<08:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4692     0.1163        125        640:  50%|████▉     | 403/809 [07:55<08:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4693     0.1163        152        640:  50%|█████     | 405/809 [07:57<07:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4694     0.1162        154        640:  50%|█████     | 406/809 [07:58<08:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4691     0.1163        378        640:  50%|█████     | 408/809 [08:01<07:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4689     0.1162        153        640:  51%|█████     | 410/809 [08:03<07:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085      0.469     0.1161        162        640:  51%|█████     | 411/809 [08:04<07:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084      0.469      0.116        151        640:  51%|█████     | 413/809 [08:07<07:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4693     0.1158        253        640:  51%|█████▏    | 415/809 [08:09<07:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084      0.469     0.1161        489        640:  51%|█████▏    | 416/809 [08:10<07:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4689     0.1161        337        640:  52%|█████▏    | 418/809 [08:12<07:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4689     0.1159        141        640:  52%|█████▏    | 420/809 [08:15<07:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4689      0.116        172        640:  52%|█████▏    | 422/809 [08:17<07:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4685     0.1168        326        640:  52%|█████▏    | 423/809 [08:18<07:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4685     0.1167        232        640:  53%|█████▎    | 425/809 [08:21<07:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4682     0.1167        233        640:  53%|█████▎    | 427/809 [08:23<07:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4681     0.1165        367        640:  53%|█████▎    | 429/809 [08:25<07:18,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4681     0.1164        208        640:  53%|█████▎    | 430/809 [08:26<07:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4682     0.1163        152        640:  53%|█████▎    | 432/809 [08:29<07:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4681     0.1162        254        640:  54%|█████▎    | 434/809 [08:31<07:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4681     0.1161        344        640:  54%|█████▍    | 435/809 [08:32<07:25,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4675     0.1163        254        640:  54%|█████▍    | 437/809 [08:35<07:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4675     0.1163        142        640:  54%|█████▍    | 439/809 [08:37<07:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4675     0.1162        221        640:  54%|█████▍    | 440/809 [08:38<07:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4675     0.1161        158        640:  55%|█████▍    | 442/809 [08:41<07:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4674      0.116        349        640:  55%|█████▍    | 444/809 [08:43<07:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4673     0.1159        286        640:  55%|█████▌    | 446/809 [08:45<07:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4678     0.1158         64        640:  55%|█████▌    | 447/809 [08:46<06:56,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4675     0.1157        452        640:  56%|█████▌    | 450/809 [08:50<06:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4676     0.1157        218        640:  56%|█████▌    | 451/809 [08:51<07:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4678     0.1155        457        640:  56%|█████▌    | 453/809 [08:53<07:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4676     0.1158        281        640:  56%|█████▌    | 455/809 [08:56<06:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4677     0.1157        208        640:  56%|█████▋    | 456/809 [08:57<06:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4675     0.1157        427        640:  57%|█████▋    | 458/809 [08:59<07:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084      0.468     0.1155        211        640:  57%|█████▋    | 460/809 [09:02<06:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4679     0.1155        315        640:  57%|█████▋    | 461/809 [09:03<06:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083      0.468     0.1154        194        640:  57%|█████▋    | 463/809 [09:05<06:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4681     0.1153        212        640:  57%|█��███▋    | 465/809 [09:08<06:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4681     0.1153        245        640:  58%|█████▊    | 466/809 [09:09<06:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4681     0.1152        254        640:  58%|█████▊    | 468/809 [09:11<06:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084      0.468     0.1156        430        640:  58%|█████▊    | 470/809 [09:14<06:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084      0.468     0.1156        375        640:  58%|█████▊    | 471/809 [09:15<06:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4681     0.1154        119        640:  58%|█████▊    | 473/809 [09:17<06:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4684     0.1152        231        640:  59%|█████▊    | 475/809 [09:20<06:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4685     0.1152        231        640:  59%|█████▉    | 477/809 [09:22<06:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4687     0.1151        113        640:  59%|█████▉    | 478/809 [09:23<06:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4689      0.115        239        640:  59%|█████▉    | 480/809 [09:25<06:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4684     0.1162        375        640:  60%|█████▉    | 482/809 [09:28<06:35,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4684     0.1164        309        640:  60%|█████▉    | 483/809 [09:29<06:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4682     0.1164        313        640:  60%|█████▉    | 485/809 [09:31<06:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085      0.468     0.1163        202        640:  60%|██████    | 487/809 [09:34<06:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4679     0.1165        466        640:  60%|██████    | 488/809 [09:35<06:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085      0.468     0.1165        375        640:  61%|██████    | 490/809 [09:37<06:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4681     0.1163        240        640:  61%|██████    | 492/809 [09:40<06:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4685     0.1162        124        640:  61%|██████    | 494/809 [09:42<06:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4684     0.1164        384        640:  61%|██████    | 495/809 [09:43<06:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4682     0.1164        298        640:  61%|██████▏   | 497/809 [09:45<06:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4682     0.1163        325        640:  62%|██████▏   | 499/809 [09:48<06:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4682     0.1163        272        640:  62%|██████▏   | 500/809 [09:49<05:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4678     0.1164        330        640:  62%|██████▏   | 502/809 [09:51<05:53,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4679     0.1163         57        640:  62%|██████▏   | 504/809 [09:54<05:51,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4679     0.1163        167        640:  63%|██████▎   | 506/809 [09:56<05:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4679     0.1162        200        640:  63%|██████▎   | 507/809 [09:57<05:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084      0.468     0.1161        221        640:  63%|██████▎   | 509/809 [09:59<05:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084      0.468     0.1161        186        640:  63%|██████▎   | 511/809 [10:02<05:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4681      0.116        182        640:  63%|██████▎   | 512/809 [10:03<05:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084      0.468     0.1159        232        640:  64%|██████▎   | 514/809 [10:05<05:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4683     0.1158         87        640:  64%|██████▍   | 516/809 [10:08<05:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4685     0.1156        127        640:  64%|██████▍   | 518/809 [10:10<05:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4687     0.1155         65        640:  64%|██████▍   | 519/809 [10:11<05:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4684     0.1157        283        640:  64%|██████▍   | 521/809 [10:14<05:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4681     0.1162        621        640:  65%|██████▍   | 523/809 [10:16<05:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4682     0.1162        112        640:  65%|██████▍   | 524/809 [10:17<05:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4679     0.1162        336        640:  65%|██████▌   | 526/809 [10:19<05:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083      0.468     0.1161        190        640:  65%|██████▌   | 528/809 [10:22<05:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4678      0.116        269        640:  65%|██████▌   | 529/809 [10:23<05:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083      0.468     0.1159        196        640:  66%|██████▌   | 531/809 [10:25<05:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4678     0.1166        192        640:  66%|██████▌   | 533/809 [10:28<05:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4679     0.1166        317        640:  66%|██████▌   | 535/809 [10:30<05:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4678     0.1165        279        640:  66%|██████▋   | 536/809 [10:31<05:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4677     0.1165        299        640:  67%|██████▋   | 538/809 [10:33<05:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4672     0.1168        361        640:  67%|██████▋   | 540/809 [10:36<05:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4673     0.1167        167        640:  67%|██████▋   | 541/809 [10:37<05:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4673     0.1166        356        640:  67%|██████▋   | 543/809 [10:39<05:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4672     0.1164        191        640:  67%|██████▋   | 545/809 [10:42<05:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4672     0.1164        211        640:  67%|██████▋   | 546/809 [10:43<05:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4671     0.1164        325        640:  68%|██████▊   | 548/809 [10:45<05:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4667     0.1166        475        640:  68%|██████▊   | 550/809 [10:48<05:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.088     0.4665     0.1166        294        640:  68%|██████▊   | 552/809 [10:50<04:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.087     0.4666     0.1165        120        640:  68%|██████▊   | 553/809 [10:51<05:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4669     0.1164         90        640:  69%|██████▊   | 555/809 [10:54<04:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086      0.467     0.1163        147        640:  69%|██████▉   | 557/809 [10:56<05:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4669     0.1163        483        640:  69%|██████▉   | 558/809 [10:57<04:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4671     0.1162        204        640:  69%|██████▉   | 560/809 [11:00<04:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4669     0.1163        515        640:  69%|██████▉   | 562/809 [11:02<04:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4669     0.1162        149        640:  70%|██████▉   | 563/809 [11:03<04:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4667     0.1162        327        640:  70%|██████▉   | 565/809 [11:05<04:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4668     0.1161        102        640:  70%|███████   | 567/809 [11:08<04:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4668      0.116         97        640:  70%|██████���   | 569/809 [11:10<04:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.086     0.4668      0.116        231        640:  70%|███████   | 570/809 [11:11<04:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4673     0.1158        123        640:  71%|███████   | 572/809 [11:14<04:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4675     0.1157        245        640:  71%|███████   | 574/809 [11:16<04:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4676     0.1156        353        640:  71%|███████   | 576/809 [11:18<04:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4675     0.1157        633        640:  71%|███████▏  | 577/809 [11:19<04:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4677     0.1156         92        640:  72%|███████▏  | 579/809 [11:22<04:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4678     0.1155        231        640:  72%|███████▏  | 581/809 [11:24<04:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4679     0.1154        315        640:  72%|███████▏  | 582/809 [11:25<04:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4677     0.1156        445        640:  72%|███████▏  | 584/809 [11:28<04:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4676     0.1157        456        640:  72%|███████▏  | 586/809 [11:30<04:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4675     0.1157        346        640:  73%|███████▎  | 587/809 [11:31<04:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4674     0.1157        552        640:  73%|███████▎  | 589/809 [11:33<04:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4673     0.1157        390        640:  73%|███████▎  | 591/809 [11:36<04:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4671     0.1157        403        640:  73%|███████▎  | 593/809 [11:38<04:27,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085      0.467     0.1157        273        640:  74%|███████▎  | 595/809 [11:41<04:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4669     0.1157        355        640:  74%|███████▎  | 596/809 [11:42<04:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4669     0.1155        120        640:  74%|███████▍  | 598/809 [11:44<04:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4671     0.1154        153        640:  74%|███████▍  | 600/809 [11:47<04:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4672     0.1154        146        640:  74%|███████▍  | 601/809 [11:48<04:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4673     0.1153        241        640:  75%|███████▍  | 603/809 [11:50<04:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4676     0.1152        134        640:  75%|███████▍  | 605/809 [11:53<04:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4677     0.1151        100        640:  75%|███████▍  | 606/809 [11:54<04:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4679     0.1151        227        640:  75%|███████▌  | 608/809 [11:56<04:05,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4682      0.115        277        640:  75%|███████▌  | 610/809 [11:59<03:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4681     0.1149        243        640:  76%|███████▌  | 611/809 [12:00<03:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083      0.468     0.1149        470        640:  76%|███████▌  | 613/809 [12:02<03:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4679     0.1149        359        640:  76%|███████▌  | 615/809 [12:04<03:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4684     0.1148         84        640:  76%|███████▋  | 617/809 [12:07<03:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4684     0.1147        209        640:  76%|███████▋  | 618/809 [12:08<03:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4683     0.1148        314        640:  77%|███████▋  | 620/809 [12:10<03:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4681     0.1147        333        640:  77%|███████▋  | 622/809 [12:13<03:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4682     0.1147        129        640:  77%|███████▋  | 623/809 [12:14<03:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4685     0.1146        179        640:  77%|███████▋  | 625/809 [12:16<03:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4686     0.1145        150        640:  78%|███████▊  | 627/809 [12:19<03:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4685     0.1144        172        640:  78%|███████▊  | 629/809 [12:21<03:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4687     0.1143        125        640:  78%|███████▊  | 630/809 [12:22<03:25,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08      0.469     0.1142        204        640:  78%|███████▊  | 632/809 [12:24<03:23,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4691     0.1141        153        640:  78%|███████▊  | 634/809 [12:27<03:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4692      0.114        193        640:  78%|███████▊  | 635/809 [12:28<03:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4692      0.114        207        640:  79%|███████▊  | 637/809 [12:30<03:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4692     0.1139        162        640:  79%|███████▉  | 639/809 [12:33<03:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4691      0.114        377        640:  79%|███████▉  | 640/809 [12:34<03:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4691     0.1139        109        640:  79%|███████▉  | 642/809 [12:36<03:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4691     0.1143        264        640:  80%|███████▉  | 644/809 [12:39<03:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4692     0.1143        298        640:  80%|███████▉  | 645/809 [12:40<03:18,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.077     0.4695     0.1142        148        640:  80%|███████▉  | 647/809 [12:42<03:18,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.077     0.4695     0.1141        234        640:  80%|████████  | 649/809 [12:45<03:14,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.077     0.4692     0.1141        395        640:  80%|████████  | 650/809 [12:46<03:14,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4692     0.1142        372        640:  81%|████████  | 652/809 [12:48<03:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078      0.469     0.1142        318        640:  81%|████████  | 654/809 [12:51<03:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4687     0.1144        386        640:  81%|████████  | 655/809 [12:52<03:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4686     0.1147        441        640:  81%|████████  | 657/809 [12:54<03:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4688     0.1146        159        640:  81%|████████▏ | 659/809 [12:57<02:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4689     0.1146        309        640:  82%|████████▏ | 661/809 [12:59<02:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079      0.469     0.1146        133        640:  82%|████████▏ | 662/809 [13:00<02:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.078     0.4692     0.1144        105        640:  82%|████████▏ | 664/809 [13:02<02:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4689     0.1146        209        640:  82%|████████▏ | 666/809 [13:05<02:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4689     0.1146        195        640:  82%|████████▏ | 667/809 [13:06<02:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4688     0.1145        198        640:  83%|████████▎ | 669/809 [13:08<02:44,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4683     0.1147        322        640:  83%|████████▎ | 671/809 [13:11<02:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082      0.468     0.1147        305        640:  83%|████████▎ | 673/809 [13:13<02:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081      0.468     0.1146        173        640:  83%|████████▎ | 674/809 [13:14<02:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081      0.468     0.1145        293        640:  84%|████████▎ | 676/809 [13:17<02:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4681     0.1144        206        640:  84%|████████▍ | 678/809 [13:19<02:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4682     0.1144        115        640:  84%|████████▍ | 679/809 [13:20<02:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4679     0.1144        229        640:  84%|████████▍ | 681/809 [13:22<02:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081      0.468     0.1143        334        640:  84%|████████▍ | 683/809 [13:25<02:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081      0.468     0.1145        250        640:  85%|████████▍ | 684/809 [13:26<02:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4681     0.1144        374        640:  85%|████████▍ | 686/809 [13:28<02:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4684     0.1143         81        640:  85%|████████▌ | 688/809 [13:31<02:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4686     0.1142        163        640:  85%|████████▌ | 690/809 [13:33<02:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4686     0.1142        157        640:  85%|████████▌ | 691/809 [13:34<02:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4685     0.1143        530        640:  86%|████████▌ | 693/809 [13:37<02:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4686     0.1142        133        640:  86%|████████▌ | 695/809 [13:39<02:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4686     0.1142        438        640:  86%|████████▌ | 696/809 [13:40<02:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4684     0.1145        308        640:  86%|████████▋ | 698/809 [13:42<02:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4682     0.1145        219        640:  87%|████████▋ | 700/809 [13:45<02:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4682     0.1145        204        640:  87%|████████▋ | 701/809 [13:46<02:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4681     0.1145        401        640:  87%|████████▋ | 703/809 [13:48<02:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08      0.468     0.1145        161        640:  87%|████████▋ | 705/809 [13:51<02:04,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4682     0.1144        161        640:  87%|████████▋ | 707/809 [13:53<02:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4681     0.1144        335        640:  88%|████████▊ | 708/809 [13:54<01:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4679     0.1145        244        640:  88%|████████▊ | 710/809 [13:57<01:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4679     0.1145        148        640:  88%|████████▊ | 712/809 [13:59<01:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081      0.468     0.1144        153        640:  88%|████████▊ | 713/809 [14:00<01:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4681     0.1144        191        640:  88%|████████▊ | 715/809 [14:02<01:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4683     0.1143        121        640:  89%|████████▊ | 717/809 [14:05<01:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4684     0.1145        107        640:  89%|████████▉ | 719/809 [14:07<01:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4682     0.1145        369        640:  89%|████████▉ | 720/809 [14:08<01:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4684     0.1144        189        640:  89%|████████▉ | 722/809 [14:11<01:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4684     0.1143        236        640:  89%|████████▉ | 724/809 [14:13<01:43,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4683     0.1144        281        640:  90%|████████▉ | 726/809 [14:16<01:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4684     0.1143        171        640:  90%|████████▉ | 727/809 [14:17<01:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4684     0.1142        184        640:  90%|█████████ | 729/809 [14:19<01:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4684     0.1142        261        640:  90%|█████████ | 731/809 [14:22<01:32,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.079     0.4683     0.1142        279        640:  90%|█████████ | 732/809 [14:23<01:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4682     0.1145        349        640:  91%|█████████ | 734/809 [14:25<01:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4681     0.1145        203        640:  91%|█████████ | 736/809 [14:27<01:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08      0.468     0.1147        518        640:  91%|█████████ | 737/809 [14:29<01:26,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4677      0.115        549        640:  91%|█████████▏| 739/809 [14:31<01:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4676      0.115        386        640:  92%|████████���▏| 741/809 [14:33<01:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4674     0.1153        624        640:  92%|█████████▏| 743/809 [14:36<01:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4673     0.1153        330        640:  92%|█████████▏| 744/809 [14:37<01:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4672     0.1154        305        640:  92%|█████████▏| 746/809 [14:39<01:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083      0.467     0.1154        289        640:  92%|█████████▏| 748/809 [14:42<01:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4672     0.1153        111        640:  93%|█████████▎| 749/809 [14:43<01:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083      0.467     0.1153        351        640:  93%|█████████▎| 751/809 [14:45<01:06,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4669     0.1153        333        640:  93%|█████████▎| 753/809 [14:47<01:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4671     0.1152        190        640:  93%|█████████▎| 755/809 [14:50<01:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4672     0.1151        246        640:  93%|█████████▎| 756/809 [14:51<01:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4672     0.1151        172        640:  94%|█████████▎| 758/809 [14:53<00:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4671     0.1151        327        640:  94%|█████████▍| 760/809 [14:56<00:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4672      0.115        101        640:  94%|█████████▍| 761/809 [14:57<00:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083      0.467      0.115        360        640:  94%|█████████▍| 763/809 [14:59<00:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084      0.467      0.115        155        640:  95%|█████████▍| 765/809 [15:02<00:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084      0.467      0.115        217        640:  95%|█████████▍| 766/809 [15:03<00:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.085     0.4667     0.1155        331        640:  95%|█████████▍| 768/809 [15:05<00:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4668     0.1154        113        640:  95%|█████████▌| 770/809 [15:07<00:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4669     0.1153        199        640:  95%|█████████▌| 772/809 [15:10<00:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.084     0.4671     0.1152        255        640:  96%|█████████▌| 773/809 [15:11<00:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4672     0.1152         98        640:  96%|█████████▌| 775/809 [15:13<00:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.083     0.4673     0.1151        144        640:  96%|█████████▌| 777/809 [15:16<00:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4674     0.1151        161        640:  96%|█████████▌| 778/809 [15:17<00:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.082     0.4674      0.115        214        640:  96%|█████████▋| 780/809 [15:19<00:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4675     0.1149        159        640:  97%|█████████▋| 782/809 [15:21<00:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4674     0.1149        314        640:  97%|█████████▋| 784/809 [15:24<00:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4673     0.1148        233        640:  97%|█████████▋| 785/809 [15:25<00:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4674     0.1147        257        640:  97%|█████████▋| 787/809 [15:27<00:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4677     0.1146        135        640:  98%|█████████▊| 789/809 [15:30<00:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4678     0.1146        248        640:  98%|█████████▊| 790/809 [15:31<00:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4678     0.1145         61        640:  98%|█████████▊| 792/809 [15:33<00:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4679     0.1145        302        640:  98%|█████████▊| 794/809 [15:36<00:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4677     0.1145        377        640:  98%|█████████▊| 795/809 [15:37<00:17,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4679     0.1145         75        640:  99%|█████████▊| 797/809 [15:39<00:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4677     0.1146        277        640:  99%|█████████▉| 799/809 [15:42<00:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4677     0.1146        315        640:  99%|█████████▉| 800/809 [15:43<00:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G       1.08     0.4677     0.1145        253        640:  99%|█████████▉| 802/809 [15:45<00:08,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4675     0.1147        188        640:  99%|█████████▉| 804/809 [15:48<00:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4674     0.1149        398        640: 100%|█████████▉| 805/809 [15:49<00:05,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     85/300      13.8G      1.081     0.4675     0.1148        235        640: 100%|█████████▉| 807/809 [15:51<00:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   0%|          | 0/69 [00:00<?, ?it/s]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  12%|█▏        | 8/69 [00:01<00:14,  4.09it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  23%|██▎       | 16/69 [00:03<00:13,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  35%|███▍      | 24/69 [00:05<00:11,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  46%|████▋     | 32/69 [00:07<00:09,  4.09it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  59%|█████▉    | 41/69 [00:10<00:06,  4.04it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  70%|██████▉   | 48/69 [00:12<00:05,  3.61it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  81%|████████  | 56/69 [00:14<00:03,  3.94it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  93%|█████████▎| 64/69 [00:16<00:01,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                   all        548      38759      0.527       0.36      0.367      0.206
-     86/300      13.2G     0.7996     0.5463    0.08445        183        640:   0%|          | 1/809 [00:01<14:59,  1.11s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.139     0.4579     0.1111        344        640:   0%|          | 3/809 [00:03<15:55,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.015     0.4877    0.09509        149        640:   1%|          | 5/809 [00:05<15:26,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.009     0.4861    0.09022        252        640:   1%|          | 6/809 [00:06<15:21,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.013     0.4898    0.08868        321        640:   1%|          | 8/809 [00:09<15:33,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.037     0.4803    0.09353        410        640:   1%|          | 10/809 [00:11<15:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.037     0.4787    0.09464        299        640:   1%|▏         | 11/809 [00:12<16:00,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.019     0.4771    0.09762        124        640:   2%|▏         | 13/809 [00:15<15:40,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.024      0.479    0.09972        276        640:   2%|▏         | 15/809 [00:17<15:28,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.045     0.4706     0.1009        180        640:   2%|▏         | 17/809 [00:19<15:34,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.042     0.4719    0.09966        126        640:   2%|▏         | 18/809 [00:21<15:24,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.063     0.4669     0.1025        243        640:   2%|▏         | 20/809 [00:23<15:09,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.068     0.4635     0.1033        409        640:   3%|▎         | 22/809 [00:25<15:20,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.071     0.4622     0.1042        202        640:   3%|▎         | 23/809 [00:26<15:15,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.074     0.4607     0.1047        294        640:   3%|▎         | 25/809 [00:29<15:10,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.048     0.4727     0.1036        105        640:   3%|▎         | 27/809 [00:31<15:05,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.038     0.4776     0.1038         83        640:   4%|▎         | 29/809 [00:33<14:57,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.038     0.4775     0.1042        195        640:   4%|▎         | 30/809 [00:35<15:11,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.029     0.4787     0.1038        135        640:   4%|▍         | 32/809 [00:37<15:21,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.028     0.4779     0.1041        149        640:   4%|▍         | 34/809 [00:39<15:14,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.047     0.4735      0.107        345        640:   4%|▍         | 35/809 [00:40<15:13,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.057     0.4704     0.1076        268        640:   5%|▍         | 37/809 [00:43<15:10,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.073     0.4661     0.1085        507        640:   5%|▍         | 39/809 [00:45<15:00,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.068     0.4667     0.1069        260        640:   5%|▌         | 41/809 [00:47<14:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.069     0.4657     0.1062        240        640:   5%|▌         | 42/809 [00:49<14:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.076     0.4613     0.1064        195        640:   5%|▌         | 44/809 [00:51<14:47,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.086     0.4585     0.1084        347        640:   6%|▌         | 46/809 [00:53<14:39,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4567      0.108        374        640:   6%|▌         | 48/809 [00:56<14:42,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4573     0.1078        115        640:   6%|▌         | 49/809 [00:57<14:29,  1.14s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4564     0.1075         98        640:   6%|▋         | 51/809 [00:59<14:47,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4575     0.1072        102        640:   7%|▋         | 53/809 [01:01<14:35,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4582     0.1072        196        640:   7%|▋         | 54/809 [01:02<14:27,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G        1.1     0.4541     0.1102        240        640:   7%|▋         | 56/809 [01:05<14:43,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4516     0.1182        171        640:   7%|▋         | 58/809 [01:07<14:40,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.106      0.452     0.1176        289        640:   7%|▋         | 60/809 [01:10<14:44,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.103     0.4535     0.1172        260        640:   8%|▊         | 61/809 [01:11<14:35,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G        1.1     0.4558     0.1166        235        640:   8%|▊         | 63/809 [01:13<14:38,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.097      0.456      0.116        262        640:   8%|▊         | 64/809 [01:15<15:40,  1.26s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.102     0.4536      0.116        241        640:   8%|▊         | 66/809 [01:17<15:05,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.095     0.4549     0.1149         91        640:   8%|▊         | 68/809 [01:19<14:37,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.096     0.4552     0.1148        252        640:   9%|▊         | 70/809 [01:22<14:32,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.098     0.4545     0.1146        332        640:   9%|▉         | 71/809 [01:23<14:21,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4525     0.1167        591        640:   9%|▉         | 73/809 [01:25<14:40,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4513     0.1165        349        640:   9%|▉         | 75/809 [01:28<14:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4512     0.1167        295        640:   9%|▉         | 76/809 [01:29<14:26,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.119     0.4498     0.1181        192        640:  10%|▉         | 78/809 [01:31<14:12,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.117     0.4501     0.1173        225        640:  10%|▉         | 80/809 [01:33<14:28,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.114     0.4505     0.1169        222        640:  10%|█         | 82/809 [01:36<14:14,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.115     0.4496     0.1168        298        640:  10%|█         | 83/809 [01:37<14:08,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.122     0.4479     0.1182        207        640:  11%|█         | 85/809 [01:39<14:25,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4477     0.1175        216        640:  11%|█         | 87/809 [01:42<14:06,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.118     0.4483     0.1164        214        640:  11%|█         | 89/809 [01:44<13:58,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.118     0.4479     0.1163        325        640:  11%|█         | 90/809 [01:45<14:11,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.117     0.4481     0.1174        353        640:  11%|█▏        | 92/809 [01:48<14:20,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.113     0.4494     0.1168        384        640:  12%|█▏        | 94/809 [01:50<13:59,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.114     0.4492     0.1169        307        640:  12%|█▏        | 95/809 [01:51<13:57,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4493     0.1163        209        640:  12%|█▏        | 97/809 [01:54<14:04,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.104     0.4517     0.1155         43        640:  12%|█▏        | 99/809 [01:56<13:54,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.101     0.4523     0.1152        190        640:  12%|█▏        | 101/809 [01:58<13:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4515      0.117        532        640:  13%|█▎        | 102/809 [01:59<13:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.099     0.4528     0.1161        149        640:  13%|█▎        | 104/809 [02:02<14:10,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.099      0.453     0.1164        333        640:  13%|█▎        | 106/809 [02:04<13:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G        1.1     0.4529     0.1164        203        640:  13%|█▎        | 107/809 [02:05<13:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.098     0.4537      0.116        163        640:  13%|█▎        | 109/809 [02:08<13:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.096     0.4545     0.1155        286        640:  14%|█▎        | 111/809 [02:10<13:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.093     0.4555     0.1151        132        640:  14%|█▍        | 113/809 [02:12<13:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4563     0.1147        427        640:  14%|█▍        | 114/809 [02:14<13:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4571     0.1143        325        640:  14%|█▍        | 116/809 [02:16<13:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.087     0.4578     0.1143        361        640:  15%|█▍        | 118/809 [02:18<13:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4573     0.1144        319        640:  15%|█▍        | 119/809 [02:19<13:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.085     0.4585     0.1138        204        640:  15%|█▍        | 121/809 [02:22<13:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.087     0.4579     0.1132        380        640:  15%|█▌        | 123/809 [02:24<13:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.085     0.4588     0.1133        148        640:  15%|█��        | 125/809 [02:26<13:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.086     0.4581     0.1134        363        640:  16%|█▌        | 126/809 [02:28<13:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.085     0.4585     0.1128        273        640:  16%|█▌        | 128/809 [02:30<13:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.086     0.4578     0.1124        247        640:  16%|█▌        | 130/809 [02:32<13:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.086     0.4581     0.1122        212        640:  16%|█▌        | 131/809 [02:34<13:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.085     0.4585     0.1115        215        640:  16%|█▋        | 133/809 [02:36<13:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4575     0.1114        289        640:  17%|█▋        | 135/809 [02:38<13:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4569     0.1114        336        640:  17%|█▋        | 136/809 [02:40<13:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09     0.4569     0.1115        142        640:  17%|█▋        | 138/809 [02:42<13:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.086     0.4577      0.111        141        640:  17%|█▋        | 140/809 [02:44<13:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.086     0.4577     0.1109        181        640:  17%|█▋        | 141/809 [02:45<13:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.084     0.4579     0.1106        220        640:  18%|█▊        | 143/809 [02:48<13:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.081     0.4589     0.1102        164        640:  18%|█▊        | 145/809 [02:50<13:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.082     0.4585     0.1101        219        640:  18%|█▊        | 146/809 [02:51<13:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.084     0.4582     0.1101        363        640:  18%|█▊        | 148/809 [02:54<12:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.081     0.4594     0.1098         87        640:  19%|█▊        | 150/809 [02:56<12:40,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.087     0.4581     0.1113        323        640:  19%|█▉        | 152/809 [02:58<13:18,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.087     0.4581     0.1111        113        640:  19%|█▉        | 153/809 [03:00<13:00,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.093     0.4565     0.1118        395        640:  19%|█▉        | 155/809 [03:02<13:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.093     0.4567     0.1116        258        640:  19%|█▉        | 157/809 [03:04<12:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.093     0.4563     0.1115        195        640:  20%|█▉        | 158/809 [03:06<12:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.092     0.4567     0.1112        205        640:  20%|█▉        | 160/809 [03:08<12:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.093     0.4565     0.1113        427        640:  20%|██        | 162/809 [03:10<12:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4572     0.1111        241        640:  20%|██        | 164/809 [03:12<12:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09     0.4576     0.1108        154        640:  20%|██        | 165/809 [03:14<12:17,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4574     0.1107        408        640:  21%|██        | 167/809 [03:16<12:17,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09     0.4578     0.1105        227        640:  21%|██        | 169/809 [03:18<12:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4579     0.1102        249        640:  21%|██        | 171/809 [03:21<12:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09     0.4585       0.11        145        640:  21%|██▏       | 172/809 [03:22<12:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.093     0.4577      0.111        583        640:  22%|██▏       | 174/809 [03:24<12:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.093     0.4581     0.1109        288        640:  22%|██▏       | 176/809 [03:26<12:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4588     0.1107        207        640:  22%|██▏       | 177/809 [03:28<12:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4591     0.1109        131        640:  22%|██▏       | 179/809 [03:30<12:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4592     0.1106        143        640:  22%|██▏       | 181/809 [03:32<12:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.087     0.4602     0.1102        264        640:  23%|██▎       | 183/809 [03:35<12:00,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4599     0.1105        470        640:  23%|██▎       | 184/809 [03:36<12:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4599     0.1104        141        640:  23%|██▎       | 186/809 [03:38<12:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.093     0.4592     0.1144        705        640:  23%|██▎       | 188/809 [03:41<12:29,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.092     0.4598     0.1142        133        640:  23%|██▎       | 189/809 [03:42<12:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09     0.4603     0.1138        234        640:  24%|██▎       | 191/809 [03:44<12:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09     0.4602     0.1138        327        640:  24%|██▍       | 193/809 [03:46<12:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4607     0.1136        234        640:  24%|██▍       | 194/809 [03:48<12:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.087     0.4613     0.1133        229        640:  24%|██▍       | 196/809 [03:50<11:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4606     0.1135        469        640:  24%|██▍       | 198/809 [03:52<11:45,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09     0.4603     0.1133        414        640:  25%|██▍       | 200/809 [03:55<12:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09       0.46     0.1133        161        640:  25%|██▍       | 201/809 [03:56<12:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4595     0.1133        472        640:  25%|██▌       | 203/809 [03:58<11:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4596     0.1132        217        640:  25%|██▌       | 205/809 [04:00<11:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.094     0.4588     0.1136        445        640:  25%|██▌       | 206/809 [04:02<11:35,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.094      0.459     0.1134        208        640:  26%|██▌       | 208/809 [04:04<11:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.096     0.4587     0.1152        430        640:  26%|██▌       | 210/809 [04:06<11:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.094     0.4591      0.115        351        640:  26%|██▌       | 212/809 [04:09<11:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.092     0.4597     0.1148        178        640:  26%|██▋       | 213/809 [04:10<11:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4604     0.1145        195        640:  27%|██▋       | 215/809 [04:12<11:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.092     0.4598     0.1156        562        640:  27%|██▋       | 217/809 [04:15<11:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4601     0.1154        113        640:  27%|██▋       | 218/809 [04:16<11:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4603     0.1154        552        640:  27%|██▋       | 220/809 [04:18<11:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.093     0.4599     0.1165        163        640:  27%|██▋       | 222/809 [04:20<11:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.094     0.4597     0.1164        273        640:  28%|██▊       | 223/809 [04:22<11:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.093     0.4599     0.1162        191        640:  28%|██▊       | 225/809 [04:24<11:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4603      0.116        222        640:  28%|██▊       | 227/809 [04:27<11:49,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09     0.4605     0.1158        211        640:  28%|██▊       | 228/809 [04:28<11:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.087     0.4612     0.1154        206        640:  28%|██▊       | 230/809 [04:30<11:34,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.085     0.4617      0.115        267        640:  29%|██▊       | 232/809 [04:32<11:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.084     0.4619     0.1148        302        640:  29%|██▉       | 233/809 [04:34<11:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.082     0.4624     0.1144        331        640:  29%|██▉       | 235/809 [04:36<11:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.082      0.462     0.1142        438        640:  29%|██▉       | 237/809 [04:38<11:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.081     0.4622     0.1139        433        640:  30%|██▉       | 239/809 [04:41<11:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.079     0.4626     0.1136        239        640:  30%|██▉       | 241/809 [04:43<11:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.079     0.4626     0.1138        408        640:  30%|██▉       | 242/809 [04:44<11:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.077     0.4633     0.1138        230        640:  30%|███       | 244/809 [04:47<10:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.078     0.4632     0.1145        302        640:  30%|███       | 246/809 [04:49<10:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.078     0.4631     0.1145        311        640:  31%|███       | 247/809 [04:50<11:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.078     0.4633     0.1143        214        640:  31%|███       | 249/809 [04:53<11:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.076     0.4637     0.1141        192        640:  31%|███       | 251/809 [04:55<11:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.078     0.4633     0.1142        361        640:  31%|███▏      | 253/809 [04:57<11:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.077     0.4633     0.1142        280        640:  31%|███▏      | 254/809 [04:59<11:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.076     0.4632      0.114        171        640:  32%|███▏      | 256/809 [05:01<11:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.077     0.4628     0.1139        246        640:  32%|███▏      | 258/809 [05:03<10:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.078     0.4626     0.1144        401        640:  32%|███▏      | 259/809 [05:04<10:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.076     0.4634     0.1141        237        640:  32%|███▏      | 261/809 [05:07<10:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.076     0.4636     0.1141        115        640:  33%|███▎      | 263/809 [05:09<10:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.076     0.4635      0.114        318        640:  33%|███▎      | 264/809 [05:10<10:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.075     0.4637     0.1138        294        640:  33%|███▎      | 266/809 [05:13<10:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.076     0.4636     0.1139        175        640:  33%|███▎      | 268/809 [05:15<10:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.079     0.4631     0.1152        505        640:  33%|███▎      | 269/809 [05:16<11:02,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.079     0.4627      0.115        206        640:  33%|███▎      | 271/809 [05:19<11:06,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.079     0.4627      0.115        164        640:  34%|███▎      | 273/809 [05:21<11:19,  1.27s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.08     0.4625     0.1149        211        640:  34%|███▍      | 274/809 [05:23<10:58,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.082     0.4617      0.115        276        640:  34%|███▍      | 276/809 [05:25<10:37,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.084     0.4614     0.1155        163        640:  34%|███▍      | 278/809 [05:27<10:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.085     0.4611     0.1156        532        640:  34%|███▍      | 279/809 [05:29<10:55,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.086     0.4605     0.1161        351        640:  35%|███▍      | 281/809 [05:31<10:58,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088       0.46     0.1169        506        640:  35%|███▍      | 282/809 [05:32<10:54,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.086     0.4608     0.1166        184        640:  35%|███▌      | 284/809 [05:35<10:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.086     0.4611     0.1165        229        640:  35%|███▌      | 286/809 [05:37<10:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.086      0.461     0.1166        288        640:  36%|███▌      | 288/809 [05:39<10:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.085     0.4613     0.1165        107        640:  36%|███▌      | 289/809 [05:40<10:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.085     0.4614     0.1163        218        640:  36%|███▌      | 291/809 [05:43<10:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.087     0.4608     0.1162        242        640:  36%|███▌      | 293/809 [05:45<09:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.087     0.4608      0.116        188        640:  36%|███▋      | 295/809 [05:47<09:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4606     0.1162        439        640:  37%|███▋      | 296/809 [05:49<10:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4608      0.116        208        640:  37%|███▋      | 298/809 [05:51<09:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4606      0.117        212        640:  37%|███▋      | 300/809 [05:53<09:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4607      0.117        358        640:  37%|███▋      | 301/809 [05:54<10:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4605      0.117        281        640:  37%|███▋      | 303/809 [05:57<09:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4606     0.1168        236        640:  38%|███▊      | 305/809 [05:59<09:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4603     0.1169        414        640:  38%|███▊      | 306/809 [06:00<09:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.092     0.4595     0.1175        476        640:  38%|███▊      | 308/809 [06:03<10:03,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4595     0.1173        138        640:  38%|███▊      | 310/809 [06:05<09:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.092     0.4594     0.1172        259        640:  39%|███▊      | 312/809 [06:07<09:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4597      0.117        289        640:  39%|███▊      | 313/809 [06:09<09:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4599     0.1171        367        640:  39%|███▉      | 315/809 [06:11<09:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4601     0.1168        166        640:  39%|███▉      | 317/809 [06:13<09:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4602     0.1167        106        640:  39%|███▉      | 318/809 [06:15<09:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4602     0.1168        489        640:  40%|███▉      | 320/809 [06:17<09:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4603     0.1175        132        640:  40%|███▉      | 322/809 [06:19<09:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4604     0.1173        168        640:  40%|███▉      | 323/809 [06:20<09:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09     0.4599     0.1216        512        640:  40%|████      | 325/809 [06:23<09:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09     0.4605     0.1213        128        640:  40%|████      | 327/809 [06:25<09:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4607     0.1212        238        640:  41%|████      | 328/809 [06:26<09:44,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4606     0.1211        139        640:  41%|████      | 330/809 [06:29<09:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.088     0.4606      0.121        375        640:  41%|████      | 332/809 [06:31<09:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.089     0.4599     0.1209        413        640:  41%|████▏     | 334/809 [06:33<09:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4595      0.121        465        640:  41%|████▏     | 335/809 [06:35<09:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.09     0.4595     0.1209        424        640:  42%|████▏     | 337/809 [06:37<09:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.092     0.4591     0.1212         78        640:  42%|████▏     | 339/809 [06:39<09:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091     0.4592     0.1211        102        640:  42%|████▏     | 340/809 [06:41<09:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.091      0.459      0.121        371        640:  42%|████▏     | 342/809 [06:43<08:59,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.092     0.4588     0.1209        361        640:  43%|████▎     | 344/809 [06:45<09:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.093     0.4585     0.1209        111        640:  43%|████▎     | 346/809 [06:48<09:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.094     0.4581     0.1211        319        640:  43%|████▎     | 347/809 [06:49<09:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.096     0.4575     0.1212        359        640:  43%|████▎     | 349/809 [06:51<08:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.095     0.4574      0.121        202        640:  43%|████▎     | 351/809 [06:53<08:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.097     0.4569     0.1215        471        640:  44%|████▎     | 352/809 [06:55<09:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.096     0.4571     0.1213        125        640:  44%|████▍     | 354/809 [06:57<08:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.097      0.457     0.1212        278        640:  44%|████▍     | 356/809 [06:59<08:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.096      0.457      0.121        216        640:  44%|████▍     | 358/809 [07:02<08:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.097      0.457     0.1211        253        640:  44%|████▍     | 359/809 [07:03<08:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.097     0.4571     0.1211        158        640:  45%|████▍     | 361/809 [07:05<08:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.098     0.4565     0.1212        577        640:  45%|████▍     | 363/809 [07:08<08:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.099     0.4563     0.1213        305        640:  45%|████▍     | 364/809 [07:09<08:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.099     0.4562     0.1211        247        640:  45%|████▌     | 366/809 [07:11<08:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.099     0.4562     0.1211        329        640:  45%|████▌     | 368/809 [07:14<08:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.099     0.4562      0.121        325        640:  46%|████▌     | 369/809 [07:15<08:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.097     0.4566     0.1208        132        640:  46%|████▌     | 371/809 [07:17<08:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.097     0.4567     0.1208        412        640:  46%|████▌     | 373/809 [07:19<08:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.096      0.457     0.1207        292        640:  46%|████▌     | 374/809 [07:21<08:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.096     0.4569     0.1206        216        640:  46%|████▋     | 376/809 [07:23<08:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.095      0.457     0.1205        132        640:  47%|████▋     | 378/809 [07:25<08:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.096     0.4568     0.1208        392        640:  47%|████▋     | 380/809 [07:28<08:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.097     0.4565     0.1213        580        640:  47%|████▋     | 381/809 [07:29<08:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.098     0.4562     0.1212        254        640:  47%|████▋     | 383/809 [07:31<08:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.097     0.4563     0.1211        195        640:  48%|████▊     | 385/809 [07:34<08:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.098     0.4559     0.1211        375        640:  48%|████▊     | 386/809 [07:35<08:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.101      0.455     0.1213        214        640:  48%|████▊     | 388/809 [07:37<08:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.104     0.4542     0.1217        438        640:  48%|████▊     | 390/809 [07:40<08:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.106     0.4535     0.1222        193        640:  48%|████▊     | 392/809 [07:42<08:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4532     0.1222        183        640:  49%|████▊     | 393/809 [07:43<08:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4527     0.1222        314        640:  49%|████▉     | 395/809 [07:45<08:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4525     0.1221        285        640:  49%|████▉     | 397/809 [07:48<08:25,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4523     0.1221        204        640:  49%|████▉     | 398/809 [07:49<08:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4518     0.1221        171        640:  49%|████▉     | 400/809 [07:52<08:16,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.114     0.4511     0.1222        269        640:  50%|████▉     | 402/809 [07:54<08:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.117     0.4504     0.1225        345        640:  50%|████▉     | 404/809 [07:56<07:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.118     0.4502     0.1226        199        640:  50%|█████     | 405/809 [07:57<07:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.118     0.4501     0.1225        211        640:  50%|█████     | 407/809 [08:00<07:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.117     0.4508     0.1224        168        640:  51%|█████     | 409/809 [08:02<07:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.118     0.4504     0.1225        272        640:  51%|█████     | 411/809 [08:04<07:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.119     0.4504     0.1225        171        640:  51%|█████     | 412/809 [08:06<07:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.117     0.4508     0.1222        140        640:  51%|█████     | 414/809 [08:08<07:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.117     0.4507     0.1222        180        640:  51%|█████▏    | 416/809 [08:10<07:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.117     0.4507     0.1222        277        640:  52%|█████▏    | 417/809 [08:11<07:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.117     0.4506     0.1221        193        640:  52%|█████▏    | 419/809 [08:14<07:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.119     0.4501     0.1224        163        640:  52%|█████▏    | 421/809 [08:16<07:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.12     0.4498     0.1223        267        640:  52%|█████▏    | 422/809 [08:17<07:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4494     0.1223        262        640:  52%|█████▏    | 424/809 [08:20<07:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.12     0.4495     0.1222        229        640:  53%|█████▎    | 426/809 [08:22<07:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4493     0.1222        344        640:  53%|█████▎    | 428/809 [08:24<07:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4494     0.1221        127        640:  53%|█████▎    | 429/809 [08:26<07:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4493     0.1219        276        640:  53%|█████▎    | 431/809 [08:28<07:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.12     0.4494     0.1217        181        640:  54%|█████▎    | 433/809 [08:30<07:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4491      0.122        485        640:  54%|█████▎    | 434/809 [08:31<07:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4493     0.1219        300        640:  54%|█████▍    | 436/809 [08:34<07:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4491      0.122        441        640:  54%|█████▍    | 438/809 [08:36<07:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.122      0.449     0.1219        318        640:  54%|█████▍    | 439/809 [08:37<07:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4492     0.1218        265        640:  55%|█████▍    | 441/809 [08:40<07:21,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.12     0.4492     0.1216        200        640:  55%|█████▍    | 443/809 [08:42<07:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.12     0.4492     0.1215        247        640:  55%|█████▌    | 445/809 [08:44<07:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4491     0.1216        352        640:  55%|█████▌    | 446/809 [08:46<07:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.122     0.4488     0.1216        567        640:  55%|█████▌    | 448/809 [08:48<07:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4489     0.1215        422        640:  56%|█████▌    | 450/809 [08:50<06:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4492     0.1214        136        640:  56%|█████▌    | 451/809 [08:51<06:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.12     0.4496     0.1213         45        640:  56%|█████▌    | 453/809 [08:54<06:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.12     0.4494     0.1212        344        640:  56%|█████▌    | 455/809 [08:56<06:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4492     0.1212        251        640:  56%|█████▋    | 457/809 [08:59<06:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4492     0.1212        236        640:  57%|█████▋    | 458/809 [09:00<06:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4492     0.1213        268        640:  57%|█████▋    | 460/809 [09:02<06:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121      0.449     0.1212        346        640:  57%|█████▋    | 462/809 [09:04<06:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4491     0.1212        307        640:  57%|█████▋    | 463/809 [09:06<06:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.122     0.4488     0.1212        405        640:  57%|█████▋    | 465/809 [09:08<06:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4489     0.1211        342        640:  58%|█████▊    | 467/809 [09:10<06:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4491     0.1211        183        640:  58%|█████▊    | 468/809 [09:11<06:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4491     0.1209         97        640:  58%|█████▊    | 470/809 [09:14<06:43,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.121     0.4491      0.121        252        640:  58%|█████▊    | 472/809 [09:16<06:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.12     0.4492     0.1209        304        640:  58%|█████▊    | 473/809 [09:17<06:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.119     0.4495     0.1208        320        640:  59%|█████▊    | 475/809 [09:20<06:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.118     0.4497     0.1206        357        640:  59%|█████▉    | 477/809 [09:22<06:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.118     0.4498     0.1208        561        640:  59%|█████▉    | 479/809 [09:24<06:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.118     0.4499     0.1208        254        640:  59%|█████▉    | 480/809 [09:26<06:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.119     0.4495     0.1212        283        640:  60%|█████▉    | 482/809 [09:28<06:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.12     0.4495     0.1211        191        640:  60%|█████▉    | 484/809 [09:30<06:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.12     0.4493     0.1212        474        640:  60%|█████▉    | 485/809 [09:32<06:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.12     0.4494     0.1212        380        640:  60%|██████    | 487/809 [09:34<06:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.118     0.4498      0.121        114        640:  60%|██████    | 489/809 [09:36<06:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.118       0.45      0.121        205        640:  61%|██████    | 490/809 [09:38<06:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.119       0.45     0.1213        455        640:  61%|██████    | 492/809 [09:40<06:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.117     0.4504     0.1212        200        640:  61%|██████    | 494/809 [09:42<06:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.116     0.4505     0.1211         71        640:  61%|██████▏   | 496/809 [09:45<06:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.116     0.4505      0.121        330        640:  61%|██████▏   | 497/809 [09:46<06:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.115     0.4506     0.1209        260        640:  62%|██████▏   | 499/809 [09:48<06:13,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.115     0.4506     0.1207        189        640:  62%|██████▏   | 501/809 [09:51<06:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.115     0.4506     0.1206        227        640:  62%|██████▏   | 502/809 [09:52<06:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.114     0.4509     0.1205        101        640:  62%|██████▏   | 504/809 [09:54<06:11,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.114     0.4508     0.1204        280        640:  63%|██████▎   | 506/809 [09:57<06:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.115     0.4506     0.1205        479        640:  63%|██████▎   | 507/809 [09:58<05:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.114     0.4508     0.1204        259        640:  63%|██████▎   | 509/809 [10:00<05:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.113      0.451     0.1202        293        640:  63%|██████▎   | 511/809 [10:02<05:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.113     0.4511     0.1201        186        640:  63%|██████▎   | 512/809 [10:04<05:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4513       0.12        168        640:  64%|██████▎   | 514/809 [10:06<05:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4515     0.1199        164        640:  64%|██████▍   | 516/809 [10:09<05:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4513     0.1199        396        640:  64%|██████▍   | 517/809 [10:10<05:52,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4514     0.1199        294        640:  64%|██████▍   | 519/809 [10:12<05:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4515     0.1198        225        640:  64%|██████▍   | 521/809 [10:14<05:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4515     0.1197        227        640:  65%|██████▍   | 522/809 [10:16<05:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4514     0.1198        305        640:  65%|██████▍   | 524/809 [10:18<05:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4515     0.1197        302        640:  65%|██████▌   | 526/809 [10:21<05:38,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4514     0.1196        149        640:  65%|██████▌   | 527/809 [10:22<05:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4517     0.1196         92        640:  65%|██████▌   | 529/809 [10:24<05:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4518     0.1195        208        640:  66%|██████▌   | 531/809 [10:26<05:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4518     0.1194        294        640:  66%|██████▌   | 533/809 [10:29<05:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4515     0.1195        392        640:  66%|██████▌   | 534/809 [10:30<05:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4514     0.1195        314        640:  66%|██████▋   | 536/809 [10:32<05:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4515     0.1194        263        640:  67%|██████▋   | 538/809 [10:35<05:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4516     0.1193        162        640:  67%|██████▋   | 540/809 [10:37<05:09,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4516     0.1192        280        640:  67%|██████▋   | 542/809 [10:39<05:03,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4517     0.1192        252        640:  67%|██████▋   | 543/809 [10:40<05:06,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4518      0.119        172        640:  67%|██████▋   | 545/809 [10:43<05:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4517     0.1191        394        640:  68%|██████▊   | 547/809 [10:45<05:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4517      0.119        238        640:  68%|██████▊   | 548/809 [10:46<04:59,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4518     0.1189        193        640:  68%|██████▊   | 550/809 [10:48<05:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4519     0.1188        227        640:  68%|██████▊   | 552/809 [10:51<05:07,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4521     0.1187        414        640:  68%|██████▊   | 554/809 [10:53<05:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4522     0.1186        249        640:  69%|██████▊   | 555/809 [10:54<05:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4524     0.1185        121        640:  69%|██████▉   | 557/809 [10:57<04:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.106     0.4526     0.1184        326        640:  69%|██████▉   | 559/809 [10:59<04:48,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4524     0.1185        290        640:  69%|██████▉   | 560/809 [11:00<04:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4523     0.1193        379        640:  69%|██████▉   | 562/809 [11:03<04:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109      0.452     0.1196        471        640:  70%|██████▉   | 564/809 [11:05<04:53,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4519     0.1196        282        640:  70%|██████▉   | 565/809 [11:06<04:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109      0.452     0.1196        443        640:  70%|███████   | 567/809 [11:09<04:50,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4518     0.1198        370        640:  70%|███████   | 569/809 [11:11<04:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4515     0.1204        553        640:  70%|███████   | 570/809 [11:12<04:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4512     0.1207        351        640:  71%|███████   | 572/809 [11:15<04:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4512     0.1207        133        640:  71%|███████   | 574/809 [11:17<04:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4511     0.1206        249        640:  71%|███████   | 575/809 [11:18<04:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4511     0.1206        362        640:  71%|███████▏  | 577/809 [11:21<04:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4511     0.1205        237        640:  72%|███████▏  | 579/809 [11:23<04:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4511     0.1206        309        640:  72%|███████▏  | 581/809 [11:25<04:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.113     0.4509     0.1208        391        640:  72%|███████▏  | 582/809 [11:27<04:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.114     0.4508     0.1208        217        640:  72%|███████▏  | 584/809 [11:29<04:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.113     0.4513     0.1207        122        640:  72%|███████▏  | 586/809 [11:31<04:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.113     0.4512     0.1216        333        640:  73%|███████▎  | 587/809 [11:32<04:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.114     0.4511     0.1216        302        640:  73%|███████▎  | 589/809 [11:35<04:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.113     0.4513     0.1215        157        640:  73%|███████▎  | 591/809 [11:37<04:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4516     0.1213        277        640:  73%|███████▎  | 593/809 [11:39<04:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.112     0.4516     0.1213        301        640:  73%|███████▎  | 594/809 [11:41<04:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4517     0.1212        399        640:  74%|███████▎  | 596/809 [11:43<04:15,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4521     0.1212        134        640:  74%|███████▍  | 598/809 [11:45<04:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4522     0.1211        203        640:  74%|███████▍  | 599/809 [11:46<04:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4523      0.121        262        640:  74%|███████▍  | 601/809 [11:49<04:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4523     0.1209        177        640:  75%|███████▍  | 603/809 [11:51<04:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4524     0.1208        210        640:  75%|███████▍  | 604/809 [11:52<04:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4526     0.1207        233        640:  75%|███████▍  | 606/809 [11:55<03:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4527     0.1206        317        640:  75%|███████▌  | 608/809 [11:57<03:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4525     0.1204        231        640:  75%|███████▌  | 610/809 [11:59<03:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4525     0.1204        128        640:  76%|███████▌  | 611/809 [12:01<03:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4522     0.1203        265        640:  76%|███████▌  | 613/809 [12:03<03:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4521     0.1204        371        640:  76%|███████▌  | 615/809 [12:05<03:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4522     0.1204        350        640:  76%|███████▌  | 616/809 [12:06<03:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4523     0.1204        274        640:  76%|███████▋  | 618/809 [12:09<03:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4522     0.1203        385        640:  77%|███████▋  | 620/809 [12:11<03:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4521     0.1204        293        640:  77%|███████▋  | 622/809 [12:13<03:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4522     0.1205        183        640:  77%|███████▋  | 623/809 [12:15<03:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4525     0.1204        104        640:  77%|███████▋  | 625/809 [12:17<03:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4521     0.1205        578        640:  78%|███████▊  | 627/809 [12:19<03:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4521     0.1204        249        640:  78%|███████▊  | 628/809 [12:20<03:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4521     0.1205         96        640:  78%|███████▊  | 630/809 [12:23<03:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11      0.452     0.1204        304        640:  78%|███████▊  | 632/809 [12:25<03:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109      0.452     0.1204        209        640:  78%|███████▊  | 633/809 [12:26<03:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4519     0.1204        234        640:  78%|███████▊  | 635/809 [12:29<03:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4517     0.1204        307        640:  79%|███████▊  | 637/809 [12:31<03:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4518     0.1203        270        640:  79%|███████▉  | 639/809 [12:33<03:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4518     0.1202        271        640:  79%|███████▉  | 640/809 [12:35<03:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4522     0.1201        143        640:  79%|███████▉  | 642/809 [12:37<03:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4524     0.1201        269        640:  80%|███████▉  | 644/809 [12:39<03:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4525     0.1201        191        640:  80%|███████▉  | 645/809 [12:41<03:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4525     0.1199        197        640:  80%|███████▉  | 647/809 [12:43<03:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4524       0.12        301        640:  80%|████████  | 649/809 [12:45<03:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4527     0.1199        150        640:  80%|████████  | 651/809 [12:48<03:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4528     0.1199        132        640:  81%|████████  | 652/809 [12:49<03:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4528     0.1198        236        640:  81%|████████  | 654/809 [12:51<03:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4531     0.1197        190        640:  81%|████████  | 656/809 [12:53<03:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4531     0.1197        283        640:  81%|████████  | 657/809 [12:55<02:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107      0.453     0.1197        166        640:  81%|████████▏ | 659/809 [12:57<02:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107      0.453     0.1196        257        640:  82%|████████▏ | 661/809 [12:59<02:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4529     0.1197        324        640:  82%|████████▏ | 663/809 [13:02<02:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4528     0.1197        387        640:  82%|████████▏ | 664/809 [13:03<02:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4525       0.12        315        640:  82%|████████▏ | 666/809 [13:05<02:48,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4521     0.1202        694        640:  83%|████████▎ | 668/809 [13:08<02:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11      0.452     0.1202        190        640:  83%|████████▎ | 669/809 [13:09<02:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4518     0.1203        127        640:  83%|████████▎ | 671/809 [13:11<02:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4518     0.1203        260        640:  83%|████████▎ | 673/809 [13:13<02:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4518     0.1201        183        640:  83%|████████▎ | 675/809 [13:16<02:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4516     0.1202        273        640:  84%|████████▎ | 676/809 [13:17<02:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.111     0.4513     0.1203        399        640:  84%|████████▍ | 678/809 [13:19<02:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4513     0.1202        229        640:  84%|████████▍ | 680/809 [13:21<02:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G       1.11     0.4515     0.1201        115        640:  84%|████████▍ | 682/809 [13:24<02:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4517       0.12        160        640:  84%|████████▍ | 683/809 [13:25<02:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4516       0.12        435        640:  85%|████████▍ | 685/809 [13:27<02:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4517       0.12        134        640:  85%|████████▍ | 687/809 [13:30<02:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4516     0.1199        271        640:  85%|████████▌ | 688/809 [13:31<02:23,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4516     0.1198        154        640:  85%|████████▌ | 690/809 [13:33<02:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4517     0.1198        206        640:  86%|████████▌ | 692/809 [13:35<02:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4518     0.1197        242        640:  86%|████████▌ | 694/809 [13:38<02:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4517     0.1198        353        640:  86%|████████▌ | 695/809 [13:39<02:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109     0.4518     0.1197        257        640:  86%|████████▌ | 697/809 [13:41<02:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4519     0.1196        158        640:  86%|████████▋ | 699/809 [13:44<02:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4519     0.1196        418        640:  87%|████████▋ | 701/809 [13:46<02:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4521     0.1195        202        640:  87%|████████▋ | 703/809 [13:48<02:01,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4522     0.1195        259        640:  87%|████████▋ | 704/809 [13:49<02:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4523     0.1196        461        640:  87%|████████▋ | 706/809 [13:52<02:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.106     0.4524     0.1196        305        640:  88%|████████▊ | 708/809 [13:54<01:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4522     0.1197        571        640:  88%|████████▊ | 709/809 [13:55<01:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4522     0.1197        183        640:  88%|████████▊ | 711/809 [13:58<01:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.106     0.4522     0.1196        242        640:  88%|████████▊ | 713/809 [14:00<01:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.106     0.4523     0.1196        211        640:  88%|████████▊ | 714/809 [14:01<01:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.106     0.4523     0.1195        178        640:  89%|████████▊ | 716/809 [14:04<01:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4519     0.1197        402        640:  89%|████████▉ | 718/809 [14:06<01:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4521     0.1196        263        640:  89%|████████▉ | 720/809 [14:08<01:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107      0.452     0.1197        299        640:  89%|████████▉ | 721/809 [14:10<01:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4522     0.1197        146        640:  89%|████████▉ | 723/809 [14:12<01:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4521     0.1196        270        640:  90%|████████▉ | 725/809 [14:14<01:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4521     0.1196        239        640:  90%|████████▉ | 726/809 [14:15<01:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.106     0.4523     0.1195        175        640:  90%|████████▉ | 728/809 [14:18<01:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.106     0.4524     0.1194        215        640:  90%|█████████ | 730/809 [14:20<01:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4527     0.1193        177        640:  90%|█████████ | 732/809 [14:22<01:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4525     0.1195        454        640:  91%|█████████ | 733/809 [14:24<01:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4527     0.1195        266        640:  91%|█████████ | 735/809 [14:26<01:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4527     0.1195        433        640:  91%|█████████ | 737/809 [14:28<01:23,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4527     0.1195        303        640:  91%|█████████ | 738/809 [14:29<01:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.104     0.4528     0.1194        102        640:  91%|█████████▏| 740/809 [14:32<01:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4528     0.1194        271        640:  92%|█████████▏| 742/809 [14:34<01:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.104      0.453     0.1194        120        640:  92%|█████████▏| 743/809 [14:35<01:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.104      0.453     0.1193        381        640:  92%|█████████▏| 745/809 [14:38<01:16,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.104     0.4531     0.1194        165        640:  92%|█████████▏| 747/809 [14:40<01:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.104     0.4531     0.1193        214        640:  92%|█████████▏| 748/809 [14:41<01:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105      0.453     0.1194        287        640:  93%|█████████▎| 750/809 [14:44<01:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4529     0.1194        196        640:  93%|█████████▎| 752/809 [14:46<01:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105      0.453     0.1194        169        640:  93%|█████████▎| 754/809 [14:48<01:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105      0.453     0.1193        404        640:  93%|█████████▎| 755/809 [14:50<01:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4529     0.1195        521        640:  94%|█████████▎| 757/809 [14:52<01:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4529     0.1195        291        640:  94%|█████████▍| 759/809 [14:54<00:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105      0.453     0.1194        185        640:  94%|█████████▍| 760/809 [14:55<00:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4531     0.1194        332        640:  94%|█████████▍| 762/809 [14:58<00:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4531     0.1193        409        640:  94%|█████████▍| 764/809 [15:00<00:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4531     0.1193        119        640:  95%|█████████▍| 766/809 [15:03<00:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4531     0.1193        198        640:  95%|█████████▍| 767/809 [15:04<00:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4528     0.1193        341        640:  95%|█████████▌| 769/809 [15:06<00:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4531     0.1192        124        640:  95%|█████████▌| 771/809 [15:08<00:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105      0.453     0.1192        254        640:  95%|█████████▌| 772/809 [15:10<00:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.104     0.4532     0.1191        149        640:  96%|█████████▌| 774/809 [15:12<00:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105      0.453     0.1193        486        640:  96%|█████████▌| 776/809 [15:14<00:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4529     0.1193        396        640:  96%|█████████▌| 777/809 [15:16<00:39,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4529     0.1192        147        640:  96%|█████████▋| 779/809 [15:18<00:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4531     0.1191        195        640:  97%|█████████▋| 781/809 [15:20<00:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.105     0.4529     0.1192        382        640:  97%|█████████▋| 782/809 [15:21<00:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.106     0.4526     0.1195        391        640:  97%|█████████▋| 784/809 [15:24<00:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4526     0.1195        288        640:  97%|█████████▋| 786/809 [15:26<00:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4524     0.1195        253        640:  97%|█████████▋| 788/809 [15:29<00:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4524     0.1195        285        640:  98%|█████████▊| 789/809 [15:30<00:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4524     0.1195        256        640:  98%|█████████▊| 791/809 [15:32<00:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.107     0.4523     0.1195        140        640:  98%|█████████▊| 793/809 [15:35<00:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4523     0.1194        251        640:  98%|█████████▊| 794/809 [15:36<00:18,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4521     0.1198        547        640:  98%|█████████▊| 796/809 [15:38<00:16,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109      0.452     0.1198        305        640:  99%|█████████▊| 798/809 [15:41<00:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109      0.452     0.1198        194        640:  99%|█████████▉| 799/809 [15:42<00:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109      0.452     0.1197        213        640:  99%|█████████▉| 801/809 [15:44<00:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.109      0.452     0.1196        143        640:  99%|█████████▉| 803/809 [15:47<00:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4522     0.1195        119        640:  99%|█████████▉| 804/809 [15:48<00:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4522     0.1196        376        640: 100%|█████████▉| 806/809 [15:50<00:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     86/300      13.3G      1.108     0.4524     0.1195        290        640: 100%|█████████▉| 808/809 [15:53<00:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   4%|▍         | 3/69 [00:00<00:16,  4.06it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  17%|█▋        | 12/69 [00:02<00:13,  4.08it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  29%|██▉       | 20/69 [00:04<00:11,  4.11it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  41%|████      | 28/69 [00:06<00:10,  4.03it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  52%|█████▏    | 36/69 [00:08<00:08,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  64%|██████▍   | 44/69 [00:10<00:06,  3.90it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  74%|███████▍  | 51/69 [00:12<00:04,  3.76it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  86%|████████▌ | 59/69 [00:14<00:02,  3.88it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  97%|█████████▋| 67/69 [00:16<00:00,  4.10it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.024     0.4176    0.09174        334        640:   0%|          | 2/809 [00:02<15:40,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4021     0.1183        394        640:   0%|          | 3/809 [00:03<15:33,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.053     0.4406     0.1078        242        640:   1%|          | 5/809 [00:05<15:29,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.059     0.4484     0.1093        248        640:   1%|          | 7/809 [00:08<15:59,  1.20s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.02     0.4587     0.1037        378        640:   1%|          | 9/809 [00:10<15:42,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.021     0.4614     0.1017        114        640:   1%|          | 10/809 [00:11<15:35,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.002     0.4652    0.09863        170        640:   1%|▏         | 12/809 [00:14<15:45,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G     0.9851      0.472    0.09749         69        640:   2%|▏         | 14/809 [00:16<15:35,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G     0.9765     0.4776    0.09747        246        640:   2%|▏         | 15/809 [00:17<15:48,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G     0.9651     0.4859    0.09645        158        640:   2%|▏         | 17/809 [00:20<15:32,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      0.978       0.48    0.09606        404        640:   2%|▏         | 19/809 [00:22<15:20,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      0.969     0.4828    0.09584        167        640:   2%|▏         | 20/809 [00:23<15:26,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G     0.9814     0.4794    0.09566        303        640:   3%|▎         | 22/809 [00:25<15:24,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.004     0.4732    0.09698        208        640:   3%|▎         | 24/809 [00:28<15:23,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.03     0.4667     0.1053        440        640:   3%|▎         | 25/809 [00:29<15:45,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.026     0.4657     0.1054        182        640:   3%|▎         | 27/809 [00:31<15:35,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.031     0.4654     0.1052        296        640:   4%|▎         | 29/809 [00:34<15:14,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.026      0.468     0.1047        147        640:   4%|▍         | 31/809 [00:36<15:33,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.031     0.4659     0.1054        314        640:   4%|▍         | 32/809 [00:37<15:21,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.033     0.4678     0.1054        114        640:   4%|▍         | 34/809 [00:40<15:11,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.055     0.4619     0.1094        284        640:   4%|▍         | 36/809 [00:42<15:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.079     0.4552     0.1128        579        640:   5%|▍         | 38/809 [00:44<15:27,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.087     0.4535     0.1136        228        640:   5%|▍         | 39/809 [00:46<15:20,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.097     0.4523      0.114        233        640:   5%|▌         | 41/809 [00:48<15:11,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.112     0.4479      0.116        309        640:   5%|▌         | 43/809 [00:50<14:54,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4464     0.1162        246        640:   6%|▌         | 45/809 [00:53<14:53,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.116     0.4477     0.1154        166        640:   6%|▌         | 46/809 [00:54<14:45,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.129     0.4442      0.118        512        640:   6%|▌         | 48/809 [00:56<14:51,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.14     0.4412     0.1184        297        640:   6%|▌         | 50/809 [00:59<14:59,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.135     0.4434      0.118        103        640:   6%|▋         | 51/809 [01:00<14:46,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.139      0.443     0.1187        257        640:   7%|▋         | 53/809 [01:02<14:33,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.131     0.4474     0.1176         95        640:   7%|▋         | 55/809 [01:04<14:34,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.128     0.4486     0.1166        198        640:   7%|▋         | 57/809 [01:07<14:32,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.136     0.4467     0.1176        339        640:   7%|▋         | 58/809 [01:08<14:38,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.148      0.443     0.1186        546        640:   7%|▋         | 60/809 [01:10<14:29,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.159     0.4407     0.1221        641        640:   8%|▊         | 62/809 [01:12<14:41,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.158     0.4412     0.1216        285        640:   8%|▊         | 63/809 [01:14<14:46,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.171      0.437      0.124        384        640:   8%|▊         | 65/809 [01:16<14:49,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.179     0.4344     0.1253        361        640:   8%|▊         | 67/809 [01:18<14:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.187     0.4316     0.1262        294        640:   9%|▊         | 69/809 [01:21<14:29,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.191     0.4301     0.1264        190        640:   9%|▊         | 70/809 [01:22<14:24,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.191     0.4293     0.1258        179        640:   9%|▉         | 72/809 [01:24<14:23,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.183      0.431     0.1252         88        640:   9%|▉         | 74/809 [01:27<14:19,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.176     0.4327     0.1245        159        640:   9%|▉         | 75/809 [01:28<14:34,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.179     0.4315     0.1254        312        640:  10%|▉         | 77/809 [01:30<14:39,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.17     0.4349     0.1243        106        640:  10%|▉         | 79/809 [01:33<14:26,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.169     0.4348     0.1238        217        640:  10%|▉         | 80/809 [01:34<14:11,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.168     0.4349     0.1236        190        640:  10%|█         | 82/809 [01:36<14:37,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.164     0.4359     0.1228        191        640:  10%|█         | 84/809 [01:39<14:18,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.159     0.4366     0.1221        212        640:  11%|█         | 86/809 [01:41<14:00,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.16     0.4366     0.1219        389        640:  11%|█         | 87/809 [01:42<14:28,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.159     0.4369     0.1214        217        640:  11%|█         | 89/809 [01:44<14:03,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.157     0.4376     0.1209        109        640:  11%|█         | 91/809 [01:47<13:59,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4377     0.1207        225        640:  11%|█▏        | 92/809 [01:48<13:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4376     0.1203        194        640:  12%|█▏        | 94/809 [01:50<14:04,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4383     0.1198        219        640:  12%|█▏        | 96/809 [01:53<14:12,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.151     0.4388     0.1197        181        640:  12%|█▏        | 97/809 [01:54<14:19,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.149     0.4395     0.1191        426        640:  12%|█▏        | 99/809 [01:56<14:08,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.155     0.4376     0.1202        311        640:  12%|█▏        | 101/809 [01:59<14:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153      0.438     0.1201        208        640:  13%|█▎        | 102/809 [02:00<13:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.151     0.4379     0.1205         71        640:  13%|█▎        | 104/809 [02:02<14:12,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.144     0.4393     0.1199        234        640:  13%|█▎        | 106/809 [02:05<13:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.143     0.4395     0.1195        190        640:  13%|█▎        | 107/809 [02:06<13:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.144     0.4387     0.1194        351        640:  13%|█▎        | 109/809 [02:08<13:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.146     0.4385     0.1201        246        640:  14%|█▎        | 111/809 [02:11<13:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.152      0.437     0.1211        377        640:  14%|█▍        | 113/809 [02:13<13:35,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4365     0.1209        182        640:  14%|█▍        | 114/809 [02:14<13:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4355     0.1217        466        640:  14%|█▍        | 116/809 [02:16<13:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4353     0.1213        296        640:  15%|█▍        | 118/809 [02:19<13:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.155     0.4358     0.1209        145        640:  15%|█▍        | 119/809 [02:20<13:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.154     0.4356     0.1206        222        640:  15%|█▍        | 121/809 [02:22<13:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.158     0.4343     0.1211        211        640:  15%|█▌        | 123/809 [02:25<13:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156      0.435     0.1205         88        640:  15%|█▌        | 125/809 [02:27<13:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4349     0.1204        156        640:  16%|█▌        | 126/809 [02:28<13:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4358       0.12        114        640:  16%|█▌        | 128/809 [02:31<13:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.152     0.4362     0.1198        144        640:  16%|█▌        | 129/809 [02:32<13:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.151     0.4358     0.1198        134        640:  16%|█▌        | 131/809 [02:34<13:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.149     0.4361     0.1197        236        640:  16%|█▋        | 133/809 [02:36<13:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4351     0.1202        304        640:  17%|█▋        | 135/809 [02:39<13:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.154     0.4351     0.1203        335        640:  17%|█▋        | 136/809 [02:40<13:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.154      0.435     0.1201        258        640:  17%|█▋        | 138/809 [02:42<13:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.154     0.4352     0.1199        342        640:  17%|█▋        | 140/809 [02:45<12:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4356     0.1195        299        640:  18%|█▊        | 142/809 [02:47<12:46,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.15     0.4365     0.1191        112        640:  18%|█▊        | 143/809 [02:48<12:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.15     0.4369     0.1198        228        640:  18%|█▊        | 145/809 [02:50<12:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.15      0.437     0.1199        246        640:  18%|█▊        | 147/809 [02:53<12:44,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.144     0.4386     0.1192        152        640:  18%|█▊        | 148/809 [02:55<12:39,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.147      0.438     0.1195        558        640:  19%|█▊        | 150/809 [02:56<12:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.145     0.4382      0.119        179        640:  19%|█▉        | 152/809 [02:59<12:55,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.15     0.4367     0.1191        249        640:  19%|█▉        | 154/809 [03:01<12:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4357     0.1192        399        640:  19%|█▉        | 155/809 [03:02<12:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.155     0.4347     0.1191        249        640:  19%|█▉        | 157/809 [03:05<12:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.161     0.4331     0.1195        292        640:  20%|█▉        | 159/809 [03:07<13:12,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.158     0.4338     0.1194        155        640:  20%|█▉        | 160/809 [03:08<13:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156      0.434      0.119        234        640:  20%|██        | 162/809 [03:11<12:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.152     0.4349     0.1184        246        640:  20%|██        | 164/809 [03:13<12:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.151      0.435     0.1182        350        640:  20%|██        | 165/809 [03:14<12:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.148     0.4354     0.1177        267        640:  21%|██        | 167/809 [03:16<12:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.149     0.4351     0.1175        225        640:  21%|██        | 169/809 [03:19<12:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.148     0.4353      0.117        200        640:  21%|██        | 171/809 [03:21<12:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.15     0.4347     0.1171        425        640:  21%|██▏       | 172/809 [03:22<12:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.149     0.4354     0.1168        268        640:  22%|██▏       | 174/809 [03:25<12:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4341     0.1176        496        640:  22%|██▏       | 176/809 [03:27<12:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.152     0.4348     0.1176        154        640:  22%|██▏       | 177/809 [03:28<12:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4343     0.1183        236        640:  22%|██▏       | 179/809 [03:30<12:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4348     0.1183        182        640:  22%|██▏       | 181/809 [03:33<12:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.157     0.4345     0.1185        228        640:  22%|██▏       | 182/809 [03:34<12:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.157     0.4348     0.1183        172        640:  23%|██▎       | 184/809 [03:36<12:33,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4353     0.1179        194        640:  23%|██▎       | 186/809 [03:39<12:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.154     0.4361     0.1176        212        640:  23%|██▎       | 188/809 [03:41<12:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.155     0.4357     0.1175        424        640:  23%|██▎       | 190/809 [03:44<12:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.155     0.4358     0.1174        252        640:  24%|██▎       | 191/809 [03:45<12:16,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.157     0.4352     0.1184        527        640:  24%|██▍       | 193/809 [03:47<12:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.157     0.4353     0.1182        258        640:  24%|██▍       | 195/809 [03:50<12:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.159     0.4349     0.1188        508        640:  24%|██▍       | 196/809 [03:51<12:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4358     0.1185        186        640:  24%|██▍       | 198/809 [03:53<11:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156      0.436     0.1183        177        640:  25%|██▍       | 200/809 [03:55<11:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4359     0.1181        254        640:  25%|██▍       | 202/809 [03:58<11:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4358     0.1181        280        640:  25%|██▌       | 203/809 [03:59<11:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.158     0.4352     0.1186        255        640:  25%|██▌       | 205/809 [04:01<11:37,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4362     0.1184        199        640:  26%|██▌       | 207/809 [04:04<11:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.155     0.4365     0.1182        128        640:  26%|██▌       | 208/809 [04:05<11:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4374     0.1179        238        640:  26%|██▌       | 210/809 [04:07<11:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.152     0.4378     0.1177        148        640:  26%|██▌       | 212/809 [04:09<11:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.151     0.4382     0.1174        106        640:  26%|██▋       | 214/809 [04:12<11:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.151     0.4381     0.1175        231        640:  27%|██▋       | 215/809 [04:13<11:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.149     0.4389     0.1174        122        640:  27%|██▋       | 217/809 [04:15<11:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.151     0.4386     0.1173        285        640:  27%|██▋       | 219/809 [04:17<11:17,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.151     0.4383     0.1179        111        640:  27%|██▋       | 221/809 [04:20<11:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.151     0.4387     0.1178         92        640:  27%|██▋       | 222/809 [04:21<11:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.151      0.439     0.1178        290        640:  28%|██▊       | 224/809 [04:23<11:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4384     0.1179        333        640:  28%|██▊       | 226/809 [04:26<11:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4385     0.1179        229        640:  28%|██▊       | 227/809 [04:27<11:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4384     0.1176        154        640:  28%|██▊       | 229/809 [04:29<11:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.155     0.4379     0.1176        241        640:  29%|██▊       | 231/809 [04:32<11:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4373     0.1179        282        640:  29%|██▉       | 233/809 [04:34<11:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.158     0.4368     0.1183        470        640:  29%|██▉       | 234/809 [04:35<11:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.158     0.4367     0.1182        416        640:  29%|██▉       | 236/809 [04:37<11:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4373     0.1181        175        640:  29%|██▉       | 238/809 [04:40<10:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4372     0.1181        294        640:  30%|██▉       | 239/809 [04:41<11:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156      0.437     0.1181        193        640:  30%|██▉       | 241/809 [04:43<11:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.158     0.4368     0.1188        513        640:  30%|███       | 243/809 [04:46<11:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.157     0.4372     0.1187        389        640:  30%|███       | 245/809 [04:48<10:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.156     0.4374     0.1187        140        640:  30%|███       | 246/809 [04:49<10:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.155     0.4382     0.1187        150        640:  31%|███       | 248/809 [04:51<11:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4391      0.119        204        640:  31%|███       | 250/809 [04:54<10:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.154     0.4389     0.1194        320        640:  31%|███       | 251/809 [04:55<10:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.152     0.4399     0.1191        122        640:  31%|███▏      | 253/809 [04:57<10:45,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4398     0.1197        576        640:  32%|███▏      | 255/809 [05:00<11:02,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.154     0.4398     0.1199        429        640:  32%|███▏      | 257/809 [05:02<10:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.154     0.4398     0.1198        244        640:  32%|███▏      | 258/809 [05:03<10:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.153     0.4401     0.1195         76        640:  32%|███▏      | 260/809 [05:05<10:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.152     0.4402     0.1194        213        640:  32%|███▏      | 262/809 [05:08<10:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.152     0.4404     0.1194        209        640:  33%|███▎      | 263/809 [05:09<10:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.152     0.4404     0.1202        391        640:  33%|███▎      | 265/809 [05:11<10:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.151     0.4406     0.1201        437        640:  33%|███▎      | 267/809 [05:14<10:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.149     0.4415     0.1199         80        640:  33%|███▎      | 268/809 [05:15<10:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.145     0.4423     0.1195        156        640:  33%|███▎      | 270/809 [05:17<10:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.145     0.4429     0.1195        231        640:  34%|███▎      | 272/809 [05:20<10:55,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.144     0.4432     0.1193        210        640:  34%|███▎      | 273/809 [05:21<10:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.145     0.4433     0.1195        430        640:  34%|███▍      | 275/809 [05:23<10:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.145     0.4433     0.1196        232        640:  34%|███▍      | 277/809 [05:26<10:55,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.145     0.4433     0.1195        251        640:  34%|███▍      | 278/809 [05:27<10:42,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.142     0.4441     0.1192        185        640:  35%|███▍      | 280/809 [05:29<10:31,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.141     0.4443      0.119        325        640:  35%|███▍      | 282/809 [05:32<10:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.142      0.444     0.1189        372        640:  35%|███▍      | 283/809 [05:33<10:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.141     0.4442     0.1188        416        640:  35%|███▌      | 285/809 [05:35<10:17,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.14     0.4445     0.1187        418        640:  35%|███▌      | 287/809 [05:38<10:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.141     0.4443     0.1191        288        640:  36%|███▌      | 289/809 [05:40<10:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.141     0.4442      0.119        216        640:  36%|███▌      | 290/809 [05:41<10:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.141      0.444      0.119        291        640:  36%|███▌      | 292/809 [05:44<10:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.143     0.4437     0.1196        513        640:  36%|███▋      | 294/809 [05:46<10:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.142     0.4438     0.1195        232        640:  36%|███▋      | 295/809 [05:47<10:20,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.142     0.4441     0.1196        374        640:  37%|███▋      | 297/809 [05:49<10:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.14     0.4448     0.1194        270        640:  37%|███▋      | 299/809 [05:52<09:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.138     0.4457     0.1191        182        640:  37%|███▋      | 301/809 [05:54<09:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.136     0.4461      0.119        123        640:  37%|███▋      | 302/809 [05:55<09:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.137      0.446     0.1193        382        640:  38%|███▊      | 304/809 [05:58<09:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.137     0.4458     0.1192        391        640:  38%|███▊      | 306/809 [06:00<09:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.137     0.4459     0.1192        115        640:  38%|███▊      | 307/809 [06:01<09:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.138     0.4453     0.1196        531        640:  38%|███▊      | 309/809 [06:04<09:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.138     0.4454     0.1195        277        640:  38%|███▊      | 311/809 [06:06<09:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.138     0.4454     0.1195        392        640:  39%|███▊      | 312/809 [06:07<09:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.138     0.4456     0.1194        157        640:  39%|███▉      | 314/809 [06:09<09:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.138     0.4455     0.1194        362        640:  39%|███▉      | 316/809 [06:12<09:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.137     0.4459     0.1193        179        640:  39%|███▉      | 318/809 [06:14<09:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.137     0.4459     0.1192        219        640:  39%|███▉      | 319/809 [06:15<09:30,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.138     0.4458     0.1194        274        640:  40%|███▉      | 321/809 [06:18<09:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.138     0.4462     0.1194        154        640:  40%|███▉      | 323/809 [06:20<09:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.138      0.446     0.1195        481        640:  40%|████      | 324/809 [06:21<09:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.137     0.4466     0.1193        185        640:  40%|████      | 326/809 [06:23<09:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.137     0.4464     0.1191        214        640:  41%|████      | 328/809 [06:26<09:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.136     0.4469      0.119        207        640:  41%|████      | 330/809 [06:28<09:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.136     0.4468     0.1189        247        640:  41%|████      | 331/809 [06:29<09:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.136     0.4469     0.1189        150        640:  41%|████      | 333/809 [06:32<09:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.136     0.4471     0.1191        126        640:  41%|████▏     | 335/809 [06:34<09:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.136     0.4472      0.119        200        640:  42%|████▏     | 336/809 [06:35<09:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.136      0.447      0.119        164        640:  42%|████▏     | 338/809 [06:37<09:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.136     0.4474     0.1188        250        640:  42%|████▏     | 340/809 [06:40<08:59,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.135     0.4478     0.1185        210        640:  42%|████▏     | 342/809 [06:42<08:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.135     0.4477     0.1184        325        640:  43%|████▎     | 344/809 [06:45<09:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.134     0.4483     0.1182        150        640:  43%|████▎     | 346/809 [06:47<08:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.134     0.4484     0.1181        168        640:  43%|████▎     | 347/809 [06:48<08:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.133     0.4488      0.118        275        640:  43%|████▎     | 349/809 [06:50<08:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.133     0.4487     0.1179        127        640:  43%|████▎     | 351/809 [06:53<08:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.133     0.4487     0.1178        145        640:  44%|████▎     | 352/809 [06:54<08:47,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.131     0.4492     0.1175         85        640:  44%|████▍     | 354/809 [06:56<08:45,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.131      0.449     0.1176        227        640:  44%|████▍     | 356/809 [06:58<08:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4489     0.1177        362        640:  44%|████▍     | 358/809 [07:01<08:42,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4492     0.1176        124        640:  44%|████▍     | 359/809 [07:02<08:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.129     0.4501     0.1174        145        640:  45%|████▍     | 361/809 [07:04<08:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.128     0.4505     0.1176        244        640:  45%|████▍     | 363/809 [07:07<08:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.128     0.4506     0.1178        161        640:  45%|████▌     | 365/809 [07:09<08:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.128     0.4507     0.1178        209        640:  45%|████▌     | 366/809 [07:10<08:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127      0.451     0.1176        208        640:  45%|████▌     | 368/809 [07:12<08:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.129     0.4505     0.1181        484        640:  46%|████▌     | 370/809 [07:15<08:29,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.128     0.4506     0.1181        299        640:  46%|████▌     | 371/809 [07:16<08:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.129     0.4505     0.1182        157        640:  46%|████▌     | 373/809 [07:18<08:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.129     0.4506     0.1181        256        640:  46%|████▋     | 375/809 [07:21<08:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127      0.451      0.118        147        640:  47%|████▋     | 377/809 [07:23<08:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4513     0.1179        122        640:  47%|████▋     | 378/809 [07:24<08:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4516     0.1177        183        640:  47%|████▋     | 380/809 [07:26<08:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4521     0.1176         71        640:  47%|████▋     | 382/809 [07:29<08:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4522     0.1176        131        640:  47%|████▋     | 383/809 [07:30<08:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4526     0.1175        156        640:  48%|████▊     | 385/809 [07:32<08:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4523     0.1179        455        640:  48%|████▊     | 387/809 [07:35<08:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123      0.453     0.1178        108        640:  48%|████▊     | 389/809 [07:37<08:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4532     0.1178        249        640:  48%|████▊     | 390/809 [07:38<08:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4535     0.1176        242        640:  48%|████▊     | 392/809 [07:40<08:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4538     0.1176        154        640:  49%|████▊     | 394/809 [07:43<08:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4537     0.1178        496        640:  49%|████▉     | 395/809 [07:44<07:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4543     0.1176        184        640:  49%|████▉     | 397/809 [07:46<08:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4545     0.1175        237        640:  49%|████▉     | 399/809 [07:49<08:04,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4545     0.1175        329        640:  50%|████▉     | 401/809 [07:51<08:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4545     0.1174        227        640:  50%|████▉     | 402/809 [07:52<08:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4545     0.1175        240        640:  50%|████▉     | 404/809 [07:55<08:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4543     0.1174        291        640:  50%|█████     | 406/809 [07:57<07:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4538     0.1185        391        640:  50%|█████     | 407/809 [07:58<08:09,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4533     0.1185        302        640:  51%|█████     | 409/809 [08:00<07:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4529     0.1187        224        640:  51%|█████     | 411/809 [08:03<07:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4524     0.1187        252        640:  51%|█████     | 413/809 [08:05<07:34,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4523     0.1186        258        640:  51%|█████     | 414/809 [08:06<07:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4522     0.1186        296        640:  51%|█████▏    | 416/809 [08:09<07:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4524     0.1186        409        640:  52%|█████▏    | 418/809 [08:11<07:45,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4524     0.1186        302        640:  52%|█████▏    | 419/809 [08:12<07:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4526     0.1185        245        640:  52%|█████▏    | 421/809 [08:14<07:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4533     0.1182        113        640:  52%|█████▏    | 423/809 [08:17<07:48,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4532     0.1182        421        640:  52%|█████▏    | 424/809 [08:18<07:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4531     0.1182        202        640:  53%|█████▎    | 426/809 [08:20<07:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4534     0.1181        185        640:  53%|█████▎    | 428/809 [08:23<07:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4534     0.1181        331        640:  53%|█████▎    | 430/809 [08:25<07:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4537     0.1181        164        640:  53%|█████▎    | 431/809 [08:26<07:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4537     0.1181        183        640:  54%|█████▎    | 433/809 [08:29<07:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4536     0.1181        188        640:  54%|█████▍    | 435/809 [08:31<07:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4536     0.1181        215        640:  54%|█████▍    | 436/809 [08:32<07:08,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4529     0.1183        307        640:  54%|█████▍    | 438/809 [08:34<07:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4524     0.1184        526        640:  54%|█████▍    | 440/809 [08:37<07:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4523     0.1184        296        640:  55%|█████▍    | 442/809 [08:39<07:02,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4524     0.1184        180        640:  55%|█████▍    | 443/809 [08:40<07:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4528     0.1182        230        640:  55%|█████▌    | 445/809 [08:43<07:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4531     0.1181        172        640:  55%|█████▌    | 447/809 [08:45<07:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4529     0.1183        375        640:  55%|█████▌    | 448/809 [08:46<07:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4528     0.1183        301        640:  56%|█████▌    | 450/809 [08:48<07:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4528     0.1183        241        640:  56%|█████▌    | 452/809 [08:51<06:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4533     0.1184        220        640:  56%|█████▌    | 454/809 [08:53<06:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4532     0.1184        438        640:  56%|█████▌    | 455/809 [08:54<06:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4534     0.1183        277        640:  56%|█████▋    | 457/809 [08:57<07:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4534     0.1182        129        640:  57%|█████▋    | 459/809 [08:59<06:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4536     0.1182        145        640:  57%|█████▋    | 460/809 [09:00<06:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4532     0.1182        286        640:  57%|█████▋    | 462/809 [09:02<06:39,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4533     0.1183        397        640:  57%|█████▋    | 464/809 [09:05<06:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4533     0.1183        169        640:  57%|█████▋    | 465/809 [09:06<06:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4536     0.1182        334        640:  58%|█████▊    | 467/809 [09:08<06:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4536     0.1181        185        640:  58%|█████▊    | 469/809 [09:11<06:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4537      0.118        189        640:  58%|█████▊    | 471/809 [09:13<06:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4537     0.1179        319        640:  58%|█████▊    | 472/809 [09:14<06:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4539     0.1179        306        640:  59%|█████▊    | 474/809 [09:17<06:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4543     0.1179        173        640:  59%|█████▉    | 476/809 [09:19<06:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.118     0.4544     0.1178        148        640:  59%|█████▉    | 477/809 [09:20<06:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.118     0.4546     0.1177        221        640:  59%|█████▉    | 479/809 [09:23<06:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.117     0.4545     0.1176        332        640:  59%|█████▉    | 481/809 [09:25<06:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.118     0.4541     0.1179        382        640:  60%|█████▉    | 483/809 [09:27<06:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4539     0.1179        292        640:  60%|█████▉    | 484/809 [09:29<06:32,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4539     0.1179        321        640:  60%|██████    | 486/809 [09:31<06:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.118     0.4541     0.1178        254        640:  60%|██████    | 488/809 [09:33<06:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4538     0.1179        408        640:  61%|██████    | 490/809 [09:36<06:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4535     0.1181        651        640:  61%|██████    | 491/809 [09:37<06:19,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4538      0.118        192        640:  61%|██████    | 493/809 [09:39<06:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4538     0.1183         88        640:  61%|██████    | 495/809 [09:42<06:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4537     0.1182        372        640:  61%|██████▏   | 496/809 [09:43<06:10,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4536     0.1182        272        640:  62%|██████▏   | 498/809 [09:45<06:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4535     0.1189        409        640:  62%|██████▏   | 500/809 [09:48<06:19,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4534     0.1189        323        640:  62%|██████▏   | 501/809 [09:49<06:10,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4535     0.1189        239        640:  62%|██████▏   | 503/809 [09:51<06:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4536     0.1189        167        640:  62%|██████▏   | 505/809 [09:54<06:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4535     0.1189        381        640:  63%|██████▎   | 506/809 [09:55<05:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4536     0.1189        295        640:  63%|██████▎   | 508/809 [09:57<06:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4535     0.1191        416        640:  63%|██████▎   | 510/809 [10:00<06:02,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4537      0.119        275        640:  63%|██████▎   | 511/809 [10:01<06:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4534     0.1202        285        640:  63%|██████▎   | 513/809 [10:03<05:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4533     0.1202        342        640:  64%|██████▎   | 515/809 [10:06<05:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4535     0.1202        110        640:  64%|██████▍   | 516/809 [10:07<05:53,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4535     0.1201        272        640:  64%|██████▍   | 518/809 [10:09<05:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4534     0.1202        533        640:  64%|██████▍   | 520/809 [10:12<05:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4534     0.1201        191        640:  64%|██████▍   | 521/809 [10:13<05:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4535     0.1201        141        640:  65%|██████▍   | 523/809 [10:15<05:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4536     0.1199        176        640:  65%|██████▍   | 525/809 [10:18<05:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4534       0.12        290        640:  65%|██████▌   | 526/809 [10:19<05:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4537       0.12        200        640:  65%|██████▌   | 528/809 [10:21<05:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4536     0.1199        160        640:  66%|██████▌   | 530/809 [10:24<05:35,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4538     0.1198        118        640:  66%|██████▌   | 531/809 [10:25<05:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4536       0.12        310        640:  66%|██████▌   | 533/809 [10:27<05:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4535       0.12        253        640:  66%|██████▌   | 535/809 [10:29<05:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4534       0.12        153        640:  66%|██████▋   | 536/809 [10:31<05:32,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4536     0.1199         63        640:  67%|██████▋   | 538/809 [10:33<05:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4536     0.1198        201        640:  67%|██████▋   | 540/809 [10:35<05:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4537     0.1197        182        640:  67%|██████▋   | 542/809 [10:38<05:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4538     0.1197        157        640:  67%|██████▋   | 543/809 [10:39<05:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4538     0.1197        265        640:  67%|██████▋   | 545/809 [10:41<05:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4537     0.1197        218        640:  68%|██████▊   | 547/809 [10:44<05:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4537     0.1197        392        640:  68%|██████▊   | 548/809 [10:45<05:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4537     0.1196        239        640:  68%|██████▊   | 550/809 [10:47<05:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.119     0.4536     0.1196        360        640:  68%|██████▊   | 552/809 [10:50<05:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.118     0.4538     0.1195        197        640:  68%|██████▊   | 554/809 [10:52<04:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.118     0.4539     0.1195        223        640:  69%|██████▊   | 555/809 [10:53<04:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4536       0.12        401        640:  69%|██████▉   | 557/809 [10:55<04:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4536     0.1205        395        640:  69%|██████▉   | 559/809 [10:58<04:58,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4535     0.1205        150        640:  69%|██████▉   | 560/809 [10:59<04:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4533     0.1206        315        640:  69%|██████▉   | 562/809 [11:01<04:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4531     0.1207        425        640:  70%|██████▉   | 564/809 [11:04<04:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4531     0.1207        273        640:  70%|██████▉   | 565/809 [11:05<04:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4527     0.1208        483        640:  70%|███████   | 567/809 [11:07<04:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4525      0.121        410        640:  70%|███████   | 569/809 [11:10<04:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4525      0.121        414        640:  71%|███████   | 571/809 [11:12<04:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4525      0.121        260        640:  71%|███████   | 572/809 [11:13<04:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4527     0.1209        187        640:  71%|███████   | 574/809 [11:15<04:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124      0.453     0.1208        135        640:  71%|███████   | 576/809 [11:18<04:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4527     0.1209        432        640:  71%|███████▏  | 577/809 [11:19<04:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4523      0.121        370        640:  72%|███████▏  | 579/809 [11:21<04:39,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4524     0.1209         96        640:  72%|███████▏  | 581/809 [11:24<04:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4523     0.1209        260        640:  72%|███████▏  | 582/809 [11:25<04:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4524     0.1208        200        640:  72%|███████▏  | 584/809 [11:27<04:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4525     0.1207        381        640:  72%|███████▏  | 586/809 [11:30<04:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4526     0.1206        281        640:  73%|███████▎  | 588/809 [11:32<04:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4527     0.1205        314        640:  73%|███████▎  | 589/809 [11:33<04:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4527     0.1209        272        640:  73%|███████▎  | 591/809 [11:36<04:24,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4527      0.121        364        640:  73%|███████▎  | 593/809 [11:38<04:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4529      0.121        171        640:  73%|███████▎  | 594/809 [11:39<04:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123      0.453     0.1209        242        640:  74%|███████▎  | 596/809 [11:41<04:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123      0.453     0.1208        223        640:  74%|███████▍  | 598/809 [11:44<04:03,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4528     0.1208        282        640:  74%|███████▍  | 600/809 [11:46<04:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4526     0.1209        323        640:  74%|███████▍  | 601/809 [11:47<04:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4522      0.121        324        640:  75%|███████▍  | 603/809 [11:49<03:57,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4523      0.121        156        640:  75%|███████▍  | 605/809 [11:52<03:54,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4522      0.121        201        640:  75%|███████▌  | 607/809 [11:54<03:55,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4523      0.121        343        640:  75%|███████▌  | 608/809 [11:55<03:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4519     0.1211        270        640:  75%|███████▌  | 610/809 [11:58<03:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4519     0.1211        190        640:  76%|███████▌  | 612/809 [12:00<03:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4519     0.1212        204        640:  76%|███████▌  | 613/809 [12:01<03:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125      0.452     0.1211        212        640:  76%|███████▌  | 615/809 [12:04<03:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4522     0.1211        163        640:  76%|███████▋  | 617/809 [12:06<03:45,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4522     0.1211        420        640:  76%|███████▋  | 618/809 [12:07<03:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4525     0.1209        149        640:  77%|███████▋  | 620/809 [12:10<03:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122      0.453     0.1208        186        640:  77%|███████▋  | 622/809 [12:12<03:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4529     0.1208        161        640:  77%|███████▋  | 623/809 [12:13<03:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4528     0.1208        339        640:  77%|███████▋  | 625/809 [12:16<03:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4527     0.1207         70        640:  78%|███████▊  | 627/809 [12:18<03:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4526     0.1207        247        640:  78%|███████▊  | 629/809 [12:20<03:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4526     0.1207        219        640:  78%|███████▊  | 630/809 [12:21<03:26,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4526     0.1207        277        640:  78%|███████▊  | 632/809 [12:24<03:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4526     0.1209        115        640:  78%|███████▊  | 634/809 [12:26<03:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4526     0.1209        236        640:  79%|███████▊  | 636/809 [12:28<03:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4528     0.1208        279        640:  79%|███████▉  | 638/809 [12:31<03:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4529     0.1207         82        640:  79%|███████▉  | 639/809 [12:32<03:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4528     0.1206        281        640:  79%|███████▉  | 641/809 [12:34<03:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121      0.453     0.1205        201        640:  79%|███████▉  | 643/809 [12:36<03:10,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4529     0.1205        407        640:  80%|███████▉  | 644/809 [12:38<03:11,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4526     0.1207        229        640:  80%|███████▉  | 646/809 [12:40<03:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4526     0.1207        380        640:  80%|████████  | 648/809 [12:42<03:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4528     0.1206        182        640:  80%|████████  | 650/809 [12:45<03:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4528     0.1206        333        640:  80%|████████  | 651/809 [12:46<03:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4529     0.1205        213        640:  81%|████████  | 653/809 [12:48<03:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.12     0.4531     0.1204        122        640:  81%|████████  | 655/809 [12:51<03:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.121     0.4527     0.1208        402        640:  81%|████████  | 656/809 [12:52<03:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4523      0.121        307        640:  81%|████████▏ | 658/809 [12:54<02:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4522      0.121        170        640:  82%|████████▏ | 660/809 [12:57<02:59,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122      0.452      0.121        294        640:  82%|████████▏ | 661/809 [12:58<02:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4518     0.1212        371        640:  82%|████████▏ | 663/809 [13:00<02:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4517     0.1213        407        640:  82%|████████▏ | 665/809 [13:03<02:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4516     0.1213        330        640:  82%|████████▏ | 666/809 [13:04<02:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4515     0.1214        344        640:  83%|████████▎ | 668/809 [13:06<02:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4515     0.1213        259        640:  83%|████████▎ | 670/809 [13:09<02:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4515     0.1213        188        640:  83%|████████▎ | 671/809 [13:10<02:46,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4516     0.1213        187        640:  83%|████████▎ | 673/809 [13:12<02:46,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4518     0.1212         91        640:  83%|████████▎ | 675/809 [13:15<02:45,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4517     0.1213        353        640:  84%|████████▎ | 676/809 [13:16<02:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4518     0.1213        148        640:  84%|████████▍ | 678/809 [13:18<02:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4516     0.1213        279        640:  84%|████████▍ | 680/809 [13:21<02:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4516     0.1213        142        640:  84%|████████▍ | 682/809 [13:23<02:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4516     0.1213        153        640:  84%|████████▍ | 683/809 [13:24<02:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4514     0.1213        267        640:  85%|████████▍ | 685/809 [13:26<02:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4514     0.1216         78        640:  85%|████████▍ | 687/809 [13:29<02:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4515     0.1215        218        640:  85%|████████▌ | 688/809 [13:30<02:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4515     0.1217        223        640:  85%|████████▌ | 690/809 [13:32<02:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4515     0.1216        319        640:  86%|████████▌ | 692/809 [13:35<02:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4515     0.1216        400        640:  86%|████████▌ | 694/809 [13:37<02:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.122     0.4515     0.1216        322        640:  86%|████████▌ | 695/809 [13:38<02:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4515     0.1216        157        640:  86%|████████▌ | 697/809 [13:40<02:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4514     0.1216        246        640:  86%|████████▋ | 699/809 [13:43<02:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.123     0.4513     0.1216        404        640:  87%|████████▋ | 700/809 [13:44<02:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124      0.451     0.1216        395        640:  87%|████████▋ | 702/809 [13:46<02:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.124     0.4509     0.1216        166        640:  87%|████████▋ | 704/809 [13:49<02:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.125     0.4506     0.1217        429        640:  87%|████████▋ | 705/809 [13:50<02:03,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.126     0.4503     0.1218        295        640:  87%|████████▋ | 707/809 [13:52<02:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127       0.45     0.1218        217        640:  88%|████████▊ | 709/809 [13:55<01:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4498     0.1219        336        640:  88%|████████▊ | 710/809 [13:56<01:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4499     0.1219        318        640:  88%|████████▊ | 712/809 [13:58<01:59,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4499     0.1219        331        640:  88%|████████▊ | 714/809 [14:01<01:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127       0.45     0.1219         93        640:  88%|████████▊ | 715/809 [14:02<01:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4498      0.122        205        640:  89%|████████▊ | 717/809 [14:04<01:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4499     0.1219        135        640:  89%|████████▉ | 719/809 [14:07<01:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127       0.45     0.1218        197        640:  89%|████████▉ | 721/809 [14:09<01:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4499     0.1218        490        640:  89%|████████▉ | 722/809 [14:10<01:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.127     0.4498     0.1218        210        640:  89%|████████▉ | 724/809 [14:12<01:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.128     0.4497      0.122        236        640:  90%|████████▉ | 726/809 [14:15<01:39,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.128     0.4498      0.122        250        640:  90%|████████▉ | 727/809 [14:16<01:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.129     0.4495     0.1223        642        640:  90%|█████████ | 729/809 [14:18<01:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.129     0.4495     0.1222        351        640:  90%|█████████ | 731/809 [14:21<01:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.129     0.4494     0.1223        320        640:  90%|█████████ | 732/809 [14:22<01:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.13     0.4492     0.1224        127        640:  91%|█████████ | 734/809 [14:24<01:31,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.13     0.4493     0.1224        373        640:  91%|█████████ | 736/809 [14:27<01:28,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.129     0.4494     0.1223        140        640:  91%|█████████ | 737/809 [14:28<01:27,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.13     0.4491     0.1228        437        640:  91%|█████████▏| 739/809 [14:31<01:28,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.13     0.4492     0.1228        501        640:  92%|█████████▏| 741/809 [14:33<01:23,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.13     0.4492     0.1227        243        640:  92%|█████████▏| 742/809 [14:34<01:21,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.131      0.449     0.1228        396        640:  92%|█████████▏| 744/809 [14:37<01:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.131     0.4488     0.1231        487        640:  92%|█████████▏| 745/809 [14:38<01:17,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.131     0.4489      0.123        234        640:  92%|█████████▏| 747/809 [14:40<01:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4488     0.1236        239        640:  93%|█████████▎| 749/809 [14:43<01:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4489     0.1235        290        640:  93%|█████████▎| 751/809 [14:45<01:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.131      0.449     0.1235        202        640:  93%|█████████▎| 752/809 [14:46<01:08,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4489     0.1235        355        640:  93%|█████████▎| 754/809 [14:49<01:05,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.131      0.449     0.1235        142        640:  93%|█████████▎| 756/809 [14:51<01:02,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4489     0.1235        242        640:  94%|█████████▎| 757/809 [14:52<01:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4487     0.1236        339        640:  94%|█████████▍| 759/809 [14:55<01:01,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4487     0.1235        248        640:  94%|█████████▍| 761/809 [14:57<00:58,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4488     0.1235        166        640:  94%|█████████▍| 762/809 [14:58<00:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4488     0.1235        159        640:  94%|█████████▍| 764/809 [15:01<00:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.133     0.4487     0.1236        371        640:  95%|█████████▍| 766/809 [15:03<00:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.133     0.4488     0.1235         61        640:  95%|█████████▍| 767/809 [15:04<00:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.133     0.4487     0.1235        260        640:  95%|█████████▌| 769/809 [15:07<00:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.133     0.4488     0.1235        299        640:  95%|█████████▌| 771/809 [15:09<00:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4489     0.1235        262        640:  96%|█████████▌| 773/809 [15:11<00:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4489     0.1234        279        640:  96%|█████████▌| 774/809 [15:12<00:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4489     0.1233        222        640:  96%|█████████▌| 776/809 [15:15<00:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4489     0.1232        295        640:  96%|█████████▌| 778/809 [15:17<00:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4487     0.1233        202        640:  96%|█████████▋| 780/809 [15:19<00:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4486     0.1232        133        640:  97%|█████████▋| 782/809 [15:22<00:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.133     0.4485     0.1232        193        640:  97%|█████████▋| 783/809 [15:23<00:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4487     0.1232        140        640:  97%|█████████▋| 785/809 [15:25<00:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4488     0.1231        298        640:  97%|█████████▋| 787/809 [15:28<00:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4487     0.1231        210        640:  97%|█████████▋| 788/809 [15:29<00:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4488      0.123        281        640:  98%|█████████▊| 790/809 [15:31<00:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4487     0.1231        244        640:  98%|█████████▊| 792/809 [15:33<00:19,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.131     0.4487      0.123        203        640:  98%|█████████▊| 794/809 [15:36<00:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4487     0.1231        468        640:  98%|█████████▊| 795/809 [15:37<00:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4487     0.1231        113        640:  99%|█████████▊| 797/809 [15:39<00:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.132     0.4487      0.123        283        640:  99%|█████████▉| 799/809 [15:42<00:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.131     0.4487      0.123        209        640:  99%|█████████▉| 800/809 [15:43<00:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.131     0.4489     0.1229        220        640:  99%|█████████▉| 802/809 [15:45<00:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.13     0.4491     0.1229        152        640:  99%|█████████▉| 804/809 [15:47<00:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G       1.13     0.4492     0.1228        278        640: 100%|█████████▉| 806/809 [15:50<00:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     87/300      13.5G      1.129     0.4493     0.1227        136        640: 100%|█████████▉| 807/809 [15:51<00:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):   3%|▎         | 2/69 [00:00<00:16,  4.04it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  14%|█▍        | 10/69 [00:02<00:14,  4.13it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  26%|██▌       | 18/69 [00:04<00:12,  4.09it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  39%|███▉      | 27/69 [00:06<00:10,  4.00it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  51%|█████     | 35/69 [00:08<00:08,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  62%|██████▏   | 43/69 [00:10<00:06,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  72%|███████▏  | 50/69 [00:12<00:04,  3.81it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  84%|████████▍ | 58/69 [00:14<00:02,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:16<00:00,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  96%|█████████▌| 66/69 [00:16<00:00,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.04it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.04it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-  0%|          | 0/809 [00:00<?, ?it/s]ances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  4.04it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.039     0.4975    0.09445        139        640:   0%|          | 2/809 [00:02<15:57,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.12     0.4613    0.09608        340        640:   0%|          | 3/809 [00:03<15:38,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G     0.9968       0.51    0.09209        102        640:   1%|          | 5/809 [00:05<15:56,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.071     0.4816     0.1089        437        640:   1%|          | 7/809 [00:08<15:50,  1.19s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.08     0.4771      0.109        422        640:   1%|          | 8/809 [00:09<15:41,  1.18s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.077     0.4788     0.1076        126        640:   1%|          | 10/809 [00:11<15:44,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.038     0.4949     0.1044        218        640:   1%|▏         | 12/809 [00:14<15:52,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.04     0.4944     0.1033        260        640:   2%|▏         | 13/809 [00:15<15:44,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.003     0.5074     0.1011         49        640:   2%|▏         | 15/809 [00:17<15:37,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.019     0.4981     0.1017        280        640:   2%|▏         | 17/809 [00:20<15:28,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.016     0.5012     0.1015        184        640:   2%|▏         | 19/809 [00:22<15:18,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.014      0.502     0.1005        126        640:   2%|▏         | 20/809 [00:23<15:23,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.043     0.4933     0.1051        368        640:   3%|▎         | 22/809 [00:25<15:28,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.064     0.4868     0.1152        546        640:   3%|▎         | 24/809 [00:28<15:21,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.061     0.4877     0.1145        242        640:   3%|▎         | 25/809 [00:29<15:13,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.076     0.4799     0.1174        386        640:   3%|▎         | 27/809 [00:31<15:21,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.08     0.4793     0.1174        214        640:   4%|▎         | 29/809 [00:34<15:11,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.064     0.4862      0.116        147        640:   4%|▍         | 31/809 [00:36<15:16,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.067     0.4856      0.117        273        640:   4%|▍         | 32/809 [00:37<15:07,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.066     0.4864     0.1155        241        640:   4%|▍         | 34/809 [00:40<15:18,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.069     0.4848     0.1155        137        640:   4%|▍         | 36/809 [00:42<15:35,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.07     0.4839     0.1153        255        640:   5%|▍         | 37/809 [00:43<15:17,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.08     0.4818     0.1196        308        640:   5%|▍         | 39/809 [00:46<15:06,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.081     0.4798      0.119        178        640:   5%|▌         | 41/809 [00:48<15:00,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.078      0.481     0.1183        183        640:   5%|▌         | 42/809 [00:49<15:32,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.074     0.4816     0.1178        118        640:   5%|▌         | 44/809 [00:51<15:05,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.077     0.4826     0.1179        134        640:   6%|▌         | 46/809 [00:54<15:11,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.088     0.4798     0.1205        522        640:   6%|▌         | 47/809 [00:55<15:01,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.093     0.4792     0.1206        410        640:   6%|▌         | 49/809 [00:57<14:49,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.098     0.4778     0.1217        497        640:   6%|▋         | 51/809 [01:00<14:40,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.094     0.4785     0.1209        316        640:   7%|▋         | 53/809 [01:02<14:34,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.089     0.4794       0.12        191        640:   7%|▋         | 54/809 [01:03<14:47,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.086     0.4797     0.1186        267        640:   7%|▋         | 56/809 [01:05<14:36,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.086     0.4792     0.1182        207        640:   7%|▋         | 58/809 [01:08<14:30,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.084       0.48     0.1182        212        640:   7%|▋         | 60/809 [01:10<14:27,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.081     0.4806     0.1178        147        640:   8%|▊         | 61/809 [01:11<14:30,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.086     0.4787     0.1175        155        640:   8%|▊         | 63/809 [01:14<14:37,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.09     0.4774     0.1174        336        640:   8%|▊         | 65/809 [01:16<14:35,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.088     0.4781     0.1168         79        640:   8%|▊         | 66/809 [01:17<14:38,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.083     0.4796     0.1155         95        640:   8%|▊         | 68/809 [01:20<14:51,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.088     0.4777     0.1156        189        640:   9%|▊         | 70/809 [01:22<14:45,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.08     0.4801     0.1148         80        640:   9%|▉         | 71/809 [01:23<15:00,  1.22s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.087     0.4772     0.1152        386        640:   9%|▉         | 73/809 [01:26<14:28,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.082     0.4773     0.1145        147        640:   9%|▉         | 75/809 [01:28<14:14,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.084     0.4758     0.1146        417        640:   9%|▉         | 76/809 [01:29<14:11,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.087     0.4742     0.1149        472        640:  10%|▉         | 78/809 [01:31<14:21,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.097      0.471     0.1154        467        640:  10%|▉         | 80/809 [01:34<14:03,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.101     0.4692     0.1153        261        640:  10%|█         | 82/809 [01:36<13:55,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.107      0.467      0.116        365        640:  10%|█         | 83/809 [01:37<13:58,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.107     0.4663     0.1158        202        640:  11%|█         | 85/809 [01:39<13:53,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.117     0.4629     0.1176        553        640:  11%|█         | 87/809 [01:42<14:04,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.116      0.463     0.1173        198        640:  11%|█         | 88/809 [01:43<14:10,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.119     0.4624     0.1185        179        640:  11%|█         | 90/809 [01:45<14:09,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.115     0.4627     0.1178        141        640:  11%|█▏        | 92/809 [01:48<14:02,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.115     0.4625     0.1178        396        640:  12%|█▏        | 94/809 [01:50<14:09,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.118     0.4617      0.118        399        640:  12%|█▏        | 95/809 [01:51<14:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.118     0.4617     0.1177        120        640:  12%|█▏        | 97/809 [01:54<13:58,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.118      0.462     0.1175        184        640:  12%|█▏        | 99/809 [01:56<13:52,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.117      0.462     0.1172        179        640:  12%|█▏        | 100/809 [01:57<13:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.12     0.4611     0.1176        282        640:  13%|█▎        | 102/809 [02:00<14:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.117     0.4622     0.1172        240        640:  13%|█▎        | 104/809 [02:02<13:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.115     0.4629     0.1169        127        640:  13%|█▎        | 106/809 [02:04<13:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.113     0.4631     0.1165        156        640:  13%|█▎        | 107/809 [02:05<13:32,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.114      0.462     0.1161        255        640:  13%|█▎        | 109/809 [02:08<13:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.11     0.4638     0.1156        129        640:  14%|█▎        | 111/809 [02:10<13:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.109     0.4642     0.1156        180        640:  14%|█▍        | 112/809 [02:11<13:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.108     0.4649     0.1153        236        640:  14%|█▍        | 114/809 [02:14<13:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.108     0.4648     0.1151        275        640:  14%|█▍        | 116/809 [02:16<13:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.107     0.4651      0.115        136        640:  14%|█▍        | 117/809 [02:17<13:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.105     0.4655     0.1148        178        640:  15%|█▍        | 119/809 [02:20<13:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.104     0.4659     0.1142        168        640:  15%|█▍        | 121/809 [02:22<13:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.104     0.4659     0.1144        216        640:  15%|█▌        | 122/809 [02:23<13:31,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.102     0.4666      0.114        253        640:  15%|█▌        | 124/809 [02:26<13:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.103     0.4662      0.115        404        640:  16%|█▌        | 126/809 [02:28<13:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.106     0.4655     0.1151        410        640:  16%|█▌        | 128/809 [02:30<13:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.106     0.4652      0.115        195        640:  16%|█▌        | 129/809 [02:31<13:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.104      0.466     0.1146         91        640:  16%|█▌        | 131/809 [02:34<13:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.104      0.466     0.1145        209        640:  16%|█▋        | 133/809 [02:36<13:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.105     0.4657     0.1146        290        640:  17%|█▋        | 135/809 [02:38<13:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.106     0.4649     0.1159        123        640:  17%|█▋        | 137/809 [02:41<13:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.106     0.4646     0.1158        394        640:  17%|█▋        | 138/809 [02:42<13:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.102     0.4654     0.1155        115        640:  17%|█▋        | 140/809 [02:44<12:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.104     0.4647     0.1157        283        640:  18%|█▊        | 142/809 [02:47<13:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.104     0.4646     0.1155        241        640:  18%|█▊        | 143/809 [02:48<13:15,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.101     0.4654     0.1152        217        640:  18%|█▊        | 145/809 [02:50<13:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.103     0.4642     0.1154        355        640:  18%|█▊        | 147/809 [02:53<13:04,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.108     0.4627     0.1168        473        640:  18%|█▊        | 149/809 [02:55<13:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.108     0.4627     0.1166        288        640:  19%|█▊        | 150/809 [02:56<13:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.112     0.4611     0.1174        277        640:  19%|█▉        | 152/809 [02:58<12:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.114     0.4602     0.1178        222        640:  19%|█▉        | 154/809 [03:01<12:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.115     0.4596     0.1182        139        640:  19%|█▉        | 155/809 [03:02<12:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.119     0.4581     0.1193        772        640:  19%|█▉        | 157/809 [03:05<13:37,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.12     0.4576     0.1195        318        640:  20%|█▉        | 158/809 [03:06<13:33,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.121     0.4571     0.1198        303        640:  20%|█▉        | 160/809 [03:08<12:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.124     0.4564     0.1205        480        640:  20%|██        | 162/809 [03:10<12:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.124     0.4559     0.1206        303        640:  20%|██        | 164/809 [03:13<12:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.127     0.4549     0.1222        521        640:  20%|██        | 165/809 [03:14<12:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.127     0.4544      0.122        196        640:  21%|██        | 167/809 [03:16<12:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.13     0.4531     0.1222        347        640:  21%|██        | 169/809 [03:19<12:45,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.129     0.4534     0.1222         88        640:  21%|██        | 170/809 [03:20<12:58,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.131     0.4524     0.1221        240        640:  21%|██▏       | 172/809 [03:23<12:47,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.134     0.4508     0.1224        373        640:  22%|██▏       | 174/809 [03:25<12:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.135     0.4507     0.1224        208        640:  22%|██▏       | 175/809 [03:26<12:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.132     0.4514     0.1221        271        640:  22%|██▏       | 177/809 [03:28<12:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.13     0.4517      0.122        126        640:  22%|██▏       | 179/809 [03:31<12:27,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.132     0.4512     0.1224        248        640:  22%|██▏       | 180/809 [03:32<12:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.133     0.4509     0.1225        202        640:  22%|██▏       | 182/809 [03:34<12:44,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.131     0.4516     0.1222        158        640:  23%|██▎       | 184/809 [03:37<12:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.132     0.4513      0.123        439        640:  23%|██▎       | 186/809 [03:39<12:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.134     0.4506     0.1233        264        640:  23%|██▎       | 187/809 [03:40<12:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.135       0.45     0.1233        288        640:  23%|██▎       | 189/809 [03:43<12:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.138     0.4495     0.1235        395        640:  24%|██▎       | 191/809 [03:45<12:09,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.138     0.4494     0.1235        285        640:  24%|██▎       | 192/809 [03:46<12:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.144     0.4476     0.1245        254        640:  24%|██▍       | 194/809 [03:49<12:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.144     0.4476     0.1243        150        640:  24%|██▍       | 196/809 [03:51<12:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.147     0.4465     0.1245        362        640:  24%|██▍       | 197/809 [03:52<12:10,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149     0.4463     0.1248        279        640:  25%|██▍       | 199/809 [03:54<12:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.148     0.4465     0.1247        165        640:  25%|██▍       | 201/809 [03:57<11:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149     0.4464     0.1247        171        640:  25%|██▌       | 203/809 [03:59<11:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.148     0.4473     0.1246         88        640:  25%|██▌       | 204/809 [04:00<11:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4462     0.1248        417        640:  25%|██▌       | 206/809 [04:03<11:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4457     0.1248        244        640:  26%|██▌       | 208/809 [04:05<11:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4456     0.1247        205        640:  26%|██▌       | 209/809 [04:06<11:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4453     0.1245        301        640:  26%|██▌       | 211/809 [04:08<11:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.147     0.4462     0.1242        147        640:  26%|██▋       | 213/809 [04:11<11:24,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.148      0.446     0.1243        279        640:  27%|██▋       | 215/809 [04:13<11:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.147     0.4463     0.1242        154        640:  27%|██▋       | 216/809 [04:14<11:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.148     0.4461     0.1256        197        640:  27%|██▋       | 218/809 [04:17<11:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.147     0.4461     0.1256        248        640:  27%|██▋       | 220/809 [04:19<11:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.148     0.4457     0.1259        385        640:  27%|██▋       | 221/809 [04:20<11:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4453     0.1261        465        640:  28%|██▊       | 223/809 [04:23<11:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4445     0.1266        251        640:  28%|██▊       | 225/809 [04:25<11:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154      0.444      0.127        249        640:  28%|██▊       | 226/809 [04:26<11:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4433     0.1273        372        640:  28%|██▊       | 228/809 [04:28<11:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4426     0.1277        173        640:  28%|██▊       | 230/809 [04:31<11:26,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4418     0.1282        258        640:  29%|██▊       | 232/809 [04:33<11:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.162     0.4416     0.1283        406        640:  29%|██▉       | 233/809 [04:34<11:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.162     0.4415     0.1282        297        640:  29%|██▉       | 235/809 [04:37<11:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4419      0.128        273        640:  29%|██▉       | 237/809 [04:39<11:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4424     0.1278        214        640:  29%|██▉       | 238/809 [04:40<11:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4422     0.1278        352        640:  30%|██▉       | 240/809 [04:43<11:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161      0.442     0.1281        554        640:  30%|██▉       | 242/809 [04:45<11:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4417     0.1281        259        640:  30%|███       | 243/809 [04:46<10:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.163     0.4415     0.1281        329        640:  30%|███       | 245/809 [04:49<11:29,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.164     0.4412     0.1281        354        640:  31%|███       | 247/809 [04:51<11:21,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.164      0.441     0.1281        304        640:  31%|███       | 248/809 [04:52<11:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.162     0.4415     0.1278        197        640:  31%|███       | 250/809 [04:55<11:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.162     0.4417     0.1275        294        640:  31%|███       | 252/809 [04:57<10:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4419     0.1274        260        640:  31%|███▏      | 253/809 [04:58<10:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4427      0.127        139        640:  32%|███▏      | 255/809 [05:01<11:05,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4426     0.1269        238        640:  32%|███▏      | 257/809 [05:03<10:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.162     0.4416     0.1277        382        640:  32%|███▏      | 259/809 [05:05<10:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.162     0.4417     0.1276        260        640:  32%|███▏      | 260/809 [05:06<10:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.164     0.4414     0.1295        443        640:  32%|███▏      | 262/809 [05:09<11:00,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.163     0.4414     0.1296        109        640:  33%|███▎      | 264/809 [05:11<10:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.162     0.4419     0.1294        145        640:  33%|███▎      | 265/809 [05:12<10:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4424     0.1292        166        640:  33%|███▎      | 267/809 [05:15<10:36,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4432     0.1288        126        640:  33%|███▎      | 269/809 [05:17<10:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4435     0.1287        240        640:  33%|███▎      | 271/809 [05:19<10:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4438     0.1286         86        640:  34%|███▎      | 272/809 [05:20<10:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4435     0.1291        268        640:  34%|███▍      | 274/809 [05:23<10:17,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4431     0.1289        290        640:  34%|███▍      | 276/809 [05:25<10:12,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4428     0.1287        292        640:  34%|███▍      | 278/809 [05:27<10:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4428     0.1285        153        640:  34%|███▍      | 279/809 [05:28<10:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155      0.443     0.1284        174        640:  35%|███▍      | 281/809 [05:31<10:10,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4432      0.128        203        640:  35%|███▍      | 283/809 [05:33<10:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154      0.443      0.128        230        640:  35%|███▌      | 284/809 [05:34<10:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4423     0.1288        574        640:  35%|███▌      | 286/809 [05:37<10:35,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4429     0.1286        305        640:  36%|███▌      | 288/809 [05:39<10:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154      0.443     0.1285        324        640:  36%|███▌      | 290/809 [05:41<10:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4436     0.1283        296        640:  36%|███▌      | 292/809 [05:44<10:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4438     0.1281        167        640:  36%|███▌      | 293/809 [05:45<10:05,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4437      0.128        315        640:  36%|███▋      | 295/809 [05:47<10:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4439     0.1277        221        640:  37%|███▋      | 297/809 [05:50<10:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4438     0.1276        297        640:  37%|███▋      | 298/809 [05:51<09:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4431      0.129        260        640:  37%|███▋      | 300/809 [05:53<10:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4426     0.1292        399        640:  37%|███▋      | 302/809 [05:56<10:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4425     0.1292        438        640:  37%|███▋      | 303/809 [05:57<10:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4424     0.1291        419        640:  38%|███▊      | 305/809 [05:59<09:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4426     0.1289        183        640:  38%|███▊      | 307/809 [06:02<09:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4427     0.1285        309        640:  38%|███▊      | 309/809 [06:04<09:39,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4424     0.1289        332        640:  38%|███▊      | 310/809 [06:05<09:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4415     0.1295        341        640:  39%|███▊      | 312/809 [06:07<09:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4413     0.1295        183        640:  39%|███▉      | 314/809 [06:10<09:37,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4412     0.1294        177        640:  39%|███▉      | 315/809 [06:11<09:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4408     0.1292        211        640:  39%|███▉      | 317/809 [06:13<09:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4407      0.129        156        640:  39%|███▉      | 319/809 [06:16<09:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16       0.44     0.1289        160        640:  40%|███▉      | 321/809 [06:18<09:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4399     0.1291        285        640:  40%|███▉      | 322/809 [06:19<09:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.164     0.4392     0.1295        232        640:  40%|████      | 324/809 [06:21<09:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.165     0.4392     0.1294        186        640:  40%|████      | 326/809 [06:24<09:40,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.166     0.4391     0.1298        301        640:  40%|████      | 327/809 [06:25<09:30,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.167     0.4388     0.1298        443        640:  41%|████      | 329/809 [06:27<09:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.166     0.4391     0.1296        121        640:  41%|████      | 331/809 [06:30<09:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.165     0.4396     0.1294        225        640:  41%|████      | 333/809 [06:32<09:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.164     0.4398     0.1293        136        640:  41%|████▏     | 334/809 [06:33<09:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.164     0.4399      0.129        254        640:  42%|████▏     | 336/809 [06:36<09:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.164     0.4397     0.1289        285        640:  42%|████▏     | 338/809 [06:38<09:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.163     0.4401     0.1288         75        640:  42%|████▏     | 339/809 [06:39<09:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.163     0.4399      0.129        179        640:  42%|████▏     | 341/809 [06:41<09:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.162     0.4404     0.1288        156        640:  42%|████▏     | 343/809 [06:44<09:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4408     0.1286        145        640:  43%|████▎     | 345/809 [06:46<09:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4408     0.1285        249        640:  43%|████▎     | 346/809 [06:47<08:56,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4408     0.1285        316        640:  43%|████▎     | 348/809 [06:50<09:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4416     0.1282        125        640:  43%|████▎     | 350/809 [06:52<08:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4417     0.1282        196        640:  43%|████▎     | 351/809 [06:53<08:53,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4417      0.128        172        640:  44%|████▎     | 353/809 [06:55<08:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4419     0.1279        195        640:  44%|████▍     | 355/809 [06:58<08:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4419     0.1278        168        640:  44%|████▍     | 357/809 [07:00<08:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4421     0.1276        142        640:  44%|████▍     | 358/809 [07:01<09:01,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4422     0.1274        238        640:  44%|████▍     | 360/809 [07:04<08:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4424     0.1273        139        640:  45%|████▍     | 362/809 [07:06<08:37,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4422     0.1273        393        640:  45%|████▍     | 363/809 [07:07<08:33,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4424     0.1272        148        640:  45%|████▌     | 365/809 [07:09<08:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4424     0.1271        207        640:  45%|████▌     | 367/809 [07:12<08:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4425     0.1271        174        640:  46%|████▌     | 369/809 [07:14<08:28,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4424     0.1272        352        640:  46%|████▌     | 370/809 [07:15<08:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4428     0.1269        118        640:  46%|████▌     | 372/809 [07:18<08:22,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4434     0.1267         85        640:  46%|████▌     | 374/809 [07:20<08:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4431     0.1265        325        640:  46%|████▋     | 376/809 [07:22<08:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4434     0.1263        227        640:  47%|████▋     | 377/809 [07:23<08:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4436     0.1262        203        640:  47%|████▋     | 379/809 [07:26<08:18,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4437     0.1259        169        640:  47%|████▋     | 381/809 [07:28<08:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4439     0.1258        174        640:  47%|████▋     | 382/809 [07:29<08:22,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4439     0.1257        317        640:  47%|████▋     | 384/809 [07:32<08:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4439     0.1255        315        640:  48%|████▊     | 386/809 [07:34<08:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4442     0.1253        101        640:  48%|████▊     | 388/809 [07:36<08:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4443     0.1252        247        640:  48%|████▊     | 389/809 [07:37<08:09,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4445     0.1251        269        640:  48%|████▊     | 391/809 [07:40<08:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149     0.4449     0.1249        166        640:  49%|████▊     | 393/809 [07:42<08:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149      0.445     0.1249        114        640:  49%|████▊     | 394/809 [07:43<08:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4445      0.125        507        640:  49%|████▉     | 396/809 [07:46<08:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4447     0.1251         65        640:  49%|████▉     | 398/809 [07:48<08:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4442     0.1253        288        640:  49%|████▉     | 400/809 [07:50<07:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4439     0.1254        293        640:  50%|████▉     | 401/809 [07:51<07:55,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4436     0.1258        439        640:  50%|████▉     | 403/809 [07:54<08:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4434     0.1259        187        640:  50%|█████     | 405/809 [07:56<07:54,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4431     0.1259        577        640:  50%|█████     | 406/809 [07:57<07:59,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4434     0.1258        148        640:  50%|█████     | 408/809 [08:00<07:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4432     0.1263        504        640:  51%|█████     | 410/809 [08:02<07:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4432     0.1261        226        640:  51%|█████     | 412/809 [08:04<07:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4433      0.126        297        640:  51%|█████     | 413/809 [08:06<07:49,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4429     0.1259        125        640:  51%|█████▏    | 415/809 [08:08<07:56,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4426      0.126        321        640:  51%|█████▏    | 416/809 [08:09<07:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4423      0.126        293        640:  52%|█████▏    | 418/809 [08:12<07:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4422      0.126        278        640:  52%|█████▏    | 420/809 [08:14<07:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4417     0.1261        290        640:  52%|█████▏    | 422/809 [08:16<07:42,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16      0.442      0.126        107        640:  52%|█████▏    | 423/809 [08:17<07:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4421     0.1261        134        640:  53%|█████▎    | 425/809 [08:20<07:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4426     0.1258        185        640:  53%|█████▎    | 427/809 [08:22<07:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4426     0.1257        294        640:  53%|█████▎    | 429/809 [08:24<07:21,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4425     0.1259        258        640:  53%|█████▎    | 430/809 [08:26<07:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4424     0.1259        178        640:  53%|█████▎    | 432/809 [08:28<07:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4424     0.1258        124        640:  54%|█████▎    | 434/809 [08:30<07:18,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4426     0.1257        157        640:  54%|█████▍    | 435/809 [08:31<07:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4429     0.1255        172        640:  54%|█████▍    | 437/809 [08:34<07:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4427     0.1257        430        640:  54%|█████▍    | 439/809 [08:36<07:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4427     0.1256        214        640:  54%|█████▍    | 440/809 [08:37<07:13,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4425     0.1257        392        640:  55%|█████▍    | 442/809 [08:40<07:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4423     0.1257        222        640:  55%|█████▍    | 444/809 [08:42<07:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4422     0.1257        286        640:  55%|█████▌    | 446/809 [08:44<07:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157      0.442     0.1257        400        640:  55%|█████▌    | 448/809 [08:47<07:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4422     0.1256         93        640:  56%|█████▌    | 449/809 [08:48<07:00,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4426     0.1254        175        640:  56%|█████▌    | 451/809 [08:50<06:53,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153      0.443     0.1253        146        640:  56%|█████▌    | 453/809 [08:52<06:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4431     0.1253        237        640:  56%|█████▌    | 455/809 [08:55<06:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4434     0.1252        316        640:  56%|█████▋    | 456/809 [08:56<06:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4433     0.1252        208        640:  57%|█████▋    | 458/809 [08:58<06:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4434     0.1252        284        640:  57%|█████▋    | 460/809 [09:01<07:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4435     0.1251        142        640:  57%|█████▋    | 461/809 [09:02<07:13,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4434      0.125        229        640:  57%|█████▋    | 463/809 [09:05<06:57,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154      0.443      0.125        222        640:  57%|█████▋    | 465/809 [09:07<06:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4431      0.125        140        640:  58%|█████▊    | 466/809 [09:08<06:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4432     0.1249         98        640:  58%|█████▊    | 468/809 [09:10<06:48,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4428      0.125        195        640:  58%|█████▊    | 470/809 [09:13<06:59,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4427     0.1249        211        640:  58%|█████▊    | 471/809 [09:14<06:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4428     0.1248        137        640:  58%|█████▊    | 473/809 [09:16<06:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4427     0.1247        286        640:  59%|█████▊    | 475/809 [09:19<06:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4424     0.1251        634        640:  59%|█████▉    | 476/809 [09:20<06:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4425      0.125        207        640:  59%|█████▉    | 478/809 [09:22<06:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4425     0.1248        196        640:  59%|█████▉    | 480/809 [09:25<06:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4424     0.1248        261        640:  60%|█████▉    | 482/809 [09:27<06:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4426     0.1247        156        640:  60%|█████▉    | 483/809 [09:28<06:23,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4429     0.1246        177        640:  60%|█████▉    | 485/809 [09:31<06:14,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4429     0.1246        141        640:  60%|██████    | 487/809 [09:33<06:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4432     0.1245        204        640:  60%|██████    | 488/809 [09:34<06:09,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15      0.443     0.1245        334        640:  61%|██████    | 490/809 [09:36<06:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15      0.443     0.1246        107        640:  61%|██████    | 492/809 [09:39<06:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149     0.4431     0.1245        119        640:  61%|██████    | 493/809 [09:40<06:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149     0.4429     0.1246        234        640:  61%|██████    | 495/809 [09:42<06:21,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4427     0.1247        237        640:  61%|██████▏   | 497/809 [09:45<06:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4426     0.1246        179        640:  62%|██████▏   | 499/809 [09:47<06:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4425     0.1246        209        640:  62%|██████▏   | 500/809 [09:48<06:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4423     0.1247        401        640:  62%|██████▏   | 502/809 [09:51<06:15,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4424     0.1246        376        640:  62%|██████▏   | 503/809 [09:52<06:20,  1.24s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4426     0.1245        263        640:  62%|██████▏   | 505/809 [09:54<06:06,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4425     0.1244        156        640:  63%|██████▎   | 507/809 [09:57<05:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4424     0.1245        239        640:  63%|██████▎   | 508/809 [09:58<05:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4424     0.1244        264        640:  63%|██████▎   | 510/809 [10:00<06:06,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149     0.4427     0.1243        117        640:  63%|██████▎   | 512/809 [10:03<05:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149     0.4427     0.1242        216        640:  64%|██████▎   | 514/809 [10:05<05:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.148     0.4429     0.1242        221        640:  64%|██████▎   | 515/809 [10:06<05:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.148     0.4429     0.1242        255        640:  64%|██████▍   | 517/809 [10:09<05:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149     0.4428     0.1242        368        640:  64%|██████▍   | 519/809 [10:11<05:41,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149     0.4427     0.1247        591        640:  64%|██████▍   | 520/809 [10:12<05:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149     0.4428     0.1246        322        640:  65%|██████▍   | 522/809 [10:15<05:47,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149     0.4429     0.1245        262        640:  65%|██████▍   | 524/809 [10:17<05:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149      0.443     0.1246        162        640:  65%|██████▍   | 525/809 [10:18<05:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.149      0.443     0.1246        165        640:  65%|██████▌   | 527/809 [10:21<05:32,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.15     0.4429     0.1246         97        640:  65%|██████▌   | 529/809 [10:23<05:26,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4425     0.1247        327        640:  66%|██████▌   | 531/809 [10:25<05:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4425     0.1247        166        640:  66%|██████▌   | 532/809 [10:26<05:19,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.151     0.4426     0.1247        162        640:  66%|██████▌   | 534/809 [10:29<05:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4424     0.1247        246        640:  66%|██████▋   | 536/809 [10:31<05:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4424     0.1246        299        640:  66%|██████▋   | 537/809 [10:32<05:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4422     0.1247        256        640:  67%|██████▋   | 539/809 [10:34<05:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4423     0.1246        274        640:  67%|██████▋   | 541/809 [10:37<05:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4421     0.1245        237        640:  67%|██████▋   | 543/809 [10:39<05:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.152     0.4421     0.1244        213        640:  67%|██████▋   | 544/809 [10:40<05:07,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4419     0.1244        143        640:  67%|██████▋   | 546/809 [10:43<05:05,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4414     0.1247        498        640:  68%|██████▊   | 548/809 [10:45<05:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4413     0.1247        305        640:  68%|██████▊   | 549/809 [10:46<05:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4412     0.1247        167        640:  68%|██████▊   | 551/809 [10:49<05:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4409     0.1246        553        640:  68%|██████▊   | 553/809 [10:51<05:08,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4407     0.1252        474        640:  68%|██████▊   | 554/809 [10:52<05:12,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4405     0.1252        377        640:  69%|██████▊   | 556/809 [10:55<04:58,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4404     0.1252        252        640:  69%|██████▉   | 558/809 [10:57<05:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4403     0.1251        253        640:  69%|██████▉   | 559/809 [10:58<04:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4401     0.1251        250        640:  69%|██████▉   | 561/809 [11:01<05:01,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4402      0.125        243        640:  70%|██████▉   | 563/809 [11:03<04:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4403     0.1249        196        640:  70%|██████▉   | 564/809 [11:04<04:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4404     0.1249        246        640:  70%|██████▉   | 566/809 [11:07<04:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4402     0.1249        468        640:  70%|███████   | 568/809 [11:09<04:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4403     0.1249        186        640:  70%|███████   | 570/809 [11:11<04:43,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4402     0.1249        176        640:  71%|███████   | 571/809 [11:13<04:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157       0.44     0.1249        399        640:  71%|███████   | 573/809 [11:15<04:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157       0.44     0.1249        340        640:  71%|███████   | 575/809 [11:17<04:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157       0.44     0.1249        199        640:  71%|███████   | 576/809 [11:18<04:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4403     0.1248        158        640:  71%|███████▏  | 578/809 [11:21<04:30,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4405     0.1248        351        640:  72%|███████▏  | 580/809 [11:23<04:25,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4405     0.1248        302        640:  72%|███████▏  | 581/809 [11:24<04:33,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4407     0.1247        233        640:  72%|███████▏  | 583/809 [11:27<04:34,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4407     0.1246        217        640:  72%|███████▏  | 585/809 [11:29<04:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4407     0.1246        272        640:  72%|███████▏  | 586/809 [11:30<04:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4409     0.1245        245        640:  73%|███████▎  | 588/809 [11:33<04:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4409     0.1244        266        640:  73%|███████▎  | 590/809 [11:35<04:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154      0.441     0.1244        141        640:  73%|███████▎  | 591/809 [11:36<04:22,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154      0.441     0.1243        153        640:  73%|███████▎  | 593/809 [11:39<04:19,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4412     0.1243        311        640:  74%|███████▎  | 595/809 [11:41<04:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4412     0.1242        182        640:  74%|███████▍  | 597/809 [11:43<04:14,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.153     0.4412     0.1242        338        640:  74%|███████▍  | 598/809 [11:45<04:15,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4409     0.1242        206        640:  74%|███████▍  | 600/809 [11:47<04:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4409     0.1242        134        640:  74%|███████▍  | 602/809 [11:49<04:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.154     0.4408     0.1241        197        640:  75%|███████▍  | 603/809 [11:51<04:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4404     0.1244        161        640:  75%|███████▍  | 605/809 [11:53<04:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4404     0.1245        473        640:  75%|███████▌  | 607/809 [11:55<04:07,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4404     0.1244        277        640:  75%|███████▌  | 609/809 [11:58<03:57,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4404     0.1244        305        640:  75%|���██████▌  | 610/809 [11:59<03:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4403     0.1244        472        640:  76%|███████▌  | 612/809 [12:01<03:56,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4402     0.1246        625        640:  76%|███████▌  | 614/809 [12:04<04:00,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4403     0.1246        198        640:  76%|███████▌  | 615/809 [12:05<03:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4403     0.1245        254        640:  76%|███████▋  | 617/809 [12:07<03:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4403     0.1245        173        640:  77%|███████▋  | 619/809 [12:10<03:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4403     0.1244        286        640:  77%|███████▋  | 621/809 [12:12<03:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4404     0.1244        234        640:  77%|███████▋  | 622/809 [12:13<03:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157       0.44     0.1247        659        640:  77%|███████▋  | 624/809 [12:16<03:39,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4396     0.1252        275        640:  77%|███████▋  | 626/809 [12:18<03:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4396     0.1253        348        640:  78%|███████▊  | 627/809 [12:19<03:39,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4398     0.1252        195        640:  78%|███████▊  | 629/809 [12:22<03:36,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4396     0.1254        218        640:  78%|███████▊  | 631/809 [12:24<03:37,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4396     0.1254        197        640:  78%|███████▊  | 632/809 [12:25<03:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4391     0.1256        340        640:  78%|███████▊  | 634/809 [12:28<03:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4389     0.1255        178        640:  79%|███████▊  | 636/809 [12:30<03:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4387     0.1255        280        640:  79%|███████▊  | 637/809 [12:31<03:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4388     0.1255        316        640:  79%|███████▉  | 639/809 [12:33<03:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161      0.439     0.1254        135        640:  79%|███████▉  | 641/809 [12:36<03:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4389     0.1253        244        640:  79%|███████▉  | 643/809 [12:38<03:13,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161      0.439     0.1253        236        640:  80%|███████▉  | 644/809 [12:39<03:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4392     0.1252        227        640:  80%|███████▉  | 646/809 [12:42<03:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161      0.439     0.1252        434        640:  80%|████████  | 648/809 [12:44<03:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4389     0.1252        335        640:  80%|████████  | 649/809 [12:45<03:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4389     0.1251        174        640:  80%|████████  | 651/809 [12:47<03:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4389     0.1251        288        640:  81%|████████  | 653/809 [12:50<02:58,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4389      0.125        326        640:  81%|████████  | 655/809 [12:52<02:58,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4388      0.125        238        640:  81%|████████  | 656/809 [12:53<02:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4389     0.1249        279        640:  81%|████████▏ | 658/809 [12:56<02:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4389     0.1249        265        640:  82%|████████▏ | 660/809 [12:58<02:56,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4389     0.1249        206        640:  82%|████████▏ | 661/809 [12:59<02:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4387     0.1249        263        640:  82%|████████▏ | 663/809 [13:01<02:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.163     0.4384     0.1257        545        640:  82%|████████▏ | 665/809 [13:04<02:52,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4387     0.1255        192        640:  82%|████████▏ | 667/809 [13:06<02:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.162     0.4385     0.1256        297        640:  83%|████████▎ | 668/809 [13:07<02:43,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4387     0.1255        219        640:  83%|████████▎ | 670/809 [13:10<02:42,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4387     0.1254        264        640:  83%|████████▎ | 672/809 [13:12<02:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4389     0.1253        147        640:  83%|████████▎ | 674/809 [13:14<02:36,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16      0.439     0.1252        138        640:  83%|████████▎ | 675/809 [13:15<02:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4389     0.1253        346        640:  84%|████████▎ | 677/809 [13:18<02:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4389     0.1253        263        640:  84%|████████▍ | 679/809 [13:20<02:33,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159      0.439     0.1252        100        640:  84%|████████▍ | 680/809 [13:21<02:31,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4389     0.1252        342        640:  84%|████████▍ | 682/809 [13:24<02:29,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4391     0.1252        220        640:  85%|████████▍ | 684/809 [13:26<02:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4391     0.1252        411        640:  85%|████████▍ | 685/809 [13:27<02:25,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4389     0.1253        370        640:  85%|████████▍ | 687/809 [13:30<02:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4389     0.1252        123        640:  85%|████████▌ | 689/809 [13:32<02:20,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4391     0.1251        200        640:  85%|████████▌ | 691/809 [13:34<02:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4389     0.1251        453        640:  86%|████████▌ | 692/809 [13:35<02:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4388     0.1251        129        640:  86%|████████▌ | 694/809 [13:38<02:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4386     0.1251        400        640:  86%|████████▌ | 696/809 [13:40<02:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4388      0.125        144        640:  86%|████████▌ | 697/809 [13:41<02:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4388      0.125        228        640:  86%|████████▋ | 699/809 [13:44<02:08,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4388     0.1253        240        640:  87%|████████▋ | 701/809 [13:46<02:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4388     0.1253        230        640:  87%|████████▋ | 702/809 [13:47<02:09,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4385     0.1254        230        640:  87%|████████▋ | 704/809 [13:50<02:06,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4384     0.1254        135        640:  87%|████████▋ | 706/809 [13:52<02:04,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4382     0.1257        429        640:  87%|████████▋ | 707/809 [13:53<02:01,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4382     0.1257        344        640:  88%|████████▊ | 709/809 [13:56<01:57,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.161     0.4381     0.1257        283        640:  88%|████████▊ | 711/809 [13:58<01:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4382     0.1256        256        640:  88%|████████▊ | 713/809 [14:00<01:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G       1.16     0.4383     0.1255        154        640:  88%|████████▊ | 714/809 [14:02<01:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4388     0.1254        114        640:  89%|████████▊ | 716/809 [14:04<01:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4387     0.1255        234        640:  89%|████████▉ | 718/809 [14:06<01:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.159     0.4388     0.1255        243        640:  89%|████████▉ | 719/809 [14:07<01:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4387     0.1255        287        640:  89%|████████▉ | 721/809 [14:10<01:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158      0.439     0.1255        119        640:  89%|████████▉ | 723/809 [14:12<01:42,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4391     0.1254        220        640:  89%|████████▉ | 724/809 [14:13<01:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4391     0.1255        234        640:  90%|████████▉ | 726/809 [14:16<01:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4391     0.1254        206        640:  90%|████████▉ | 728/809 [14:18<01:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4393     0.1254        278        640:  90%|█████████ | 730/809 [14:20<01:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4393     0.1254        324        640:  90%|█████████ | 731/809 [14:21<01:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4394     0.1254        193        640:  91%|█████████ | 733/809 [14:24<01:26,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4392     0.1257        402        640:  91%|█████████ | 735/809 [14:26<01:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4391     0.1257        293        640:  91%|█████████ | 736/809 [14:27<01:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4389     0.1261        279        640:  91%|█████████ | 738/809 [14:30<01:24,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4389      0.126        266        640:  91%|█████████▏| 740/809 [14:32<01:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158      0.439      0.126        227        640:  92%|█████████▏| 742/809 [14:34<01:20,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4392     0.1259        124        640:  92%|█████████▏| 743/809 [14:36<01:18,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4393     0.1258        400        640:  92%|█████████▏| 745/809 [14:38<01:14,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4392     0.1259        441        640:  92%|█████████▏| 747/809 [14:40<01:11,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4394     0.1259        162        640:  93%|█████████▎| 749/809 [14:42<01:08,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4392     0.1259        522        640:  93%|█████████▎| 750/809 [14:44<01:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4392     0.1258        167        640:  93%|█████████▎| 752/809 [14:46<01:05,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4393     0.1258        163        640:  93%|█████████▎| 754/809 [14:48<01:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4395     0.1257        190        640:  93%|█████████▎| 755/809 [14:49<01:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156       0.44     0.1256        193        640:  94%|█████████▎| 757/809 [14:52<01:00,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4402     0.1256        345        640:  94%|█████████▍| 759/809 [14:54<00:58,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4402     0.1255        221        640:  94%|█████████▍| 761/809 [14:56<00:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4404     0.1255        215        640:  94%|█████████▍| 763/809 [14:59<00:55,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4403     0.1254        348        640:  94%|█████████▍| 764/809 [15:00<00:53,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4403     0.1255        389        640:  95%|█████████▍| 766/809 [15:03<00:53,  1.25s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4406     0.1254        138        640:  95%|█████████▍| 768/809 [15:05<00:49,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.155     0.4406     0.1253        242        640:  95%|█████████▌| 769/809 [15:06<00:47,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4404     0.1254        266        640:  95%|█████████▌| 771/809 [15:09<00:45,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4404     0.1254        158        640:  96%|█████████▌| 773/809 [15:11<00:43,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4402     0.1256        482        640:  96%|█████████▌| 774/809 [15:12<00:43,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4404     0.1254        214        640:  96%|█████████▌| 776/809 [15:15<00:40,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4406     0.1254         65        640:  96%|█████████▌| 777/809 [15:16<00:38,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4405      0.126        272        640:  96%|█████████▋| 779/809 [15:18<00:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4407     0.1259        176        640:  97%|█████████▋| 781/809 [15:21<00:32,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4407     0.1261        143        640:  97%|█████████▋| 783/809 [15:23<00:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4406     0.1262        616        640:  97%|█████████▋| 784/809 [15:24<00:29,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4405     0.1265        450        640:  97%|█████████▋| 786/809 [15:27<00:27,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4405     0.1264        241        640:  97%|█████████▋| 788/809 [15:29<00:24,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4404     0.1264        397        640:  98%|█████████▊| 789/809 [15:30<00:23,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4407     0.1263        202        640:  98%|█████████▊| 791/809 [15:32<00:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4408     0.1262        181        640:  98%|█████████▊| 793/809 [15:35<00:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4409     0.1262        205        640:  98%|█████████▊| 795/809 [15:37<00:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4409     0.1261        368        640:  98%|█████████▊| 796/809 [15:38<00:15,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4409     0.1261        354        640:  99%|█████████▊| 798/809 [15:41<00:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.156     0.4409     0.1261        150        640:  99%|█████████▉| 800/809 [15:43<00:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4407     0.1263        528        640:  99%|█████████▉| 801/809 [15:44<00:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.158     0.4405     0.1264        474        640:  99%|█████████▉| 803/809 [15:46<00:06,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4406     0.1263        190        640: 100%|█████████▉| 805/809 [15:49<00:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4407     0.1262        200        640: 100%|█████████▉| 807/809 [15:51<00:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     88/300      13.8G      1.157     0.4407     0.1262        268        640: 100%|█████████▉| 808/809 [15:52<00:01,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  10%|█         | 7/69 [00:01<00:14,  4.17it/s]_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  22%|██▏       | 15/69 [00:03<00:13,  4.02it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  33%|███▎      | 23/69 [00:05<00:11,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  45%|████▍     | 31/69 [00:07<00:09,  4.09it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  57%|█████▋    | 39/69 [00:09<00:07,  4.05it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  67%|██████▋   | 46/69 [00:11<00:06,  3.57it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  78%|███████▊  | 54/69 [00:13<00:03,  3.89it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95):  90%|████████▉ | 62/69 [00:15<00:01,  3.98it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                 Class     Images  Instances      Box(P          R      mAP50  mAP50-95): 100%|██████████| 69/69 [00:17<00:00,  3.99it/s]2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-                   all        548      38759       0.52      0.355      0.357      0.202
-     89/300      13.6G     0.7913      0.551    0.07973         91        640:   0%|          | 1/809 [00:01<15:45,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G     0.8188     0.5228    0.09343         99        640:   0%|          | 2/809 [00:02<15:43,  1.17s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.009     0.4669     0.1002        340        640:   0%|          | 4/809 [00:04<15:33,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.115     0.4362     0.1103        306        640:   1%|          | 6/809 [00:06<15:31,  1.16s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G        1.1     0.4442     0.1079        187        640:   1%|          | 8/809 [00:09<15:23,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.163      0.428     0.1245        298        640:   1%|          | 9/809 [00:10<15:19,  1.15s/it]id_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.171     0.4274     0.1217        239        640:   1%|▏         | 11/809 [00:12<15:41,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.133     0.4349     0.1176        217        640:   2%|▏         | 13/809 [00:15<15:51,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.166     0.4265     0.1198        208        640:   2%|▏         | 14/809 [00:16<15:47,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.186     0.4225     0.1212         80        640:   2%|▏         | 16/809 [00:18<15:33,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.212     0.4139     0.1221        325        640:   2%|▏         | 18/809 [00:21<15:32,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.24     0.4057     0.1251        253        640:   2%|▏         | 19/809 [00:22<15:41,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.267     0.3981     0.1257        173        640:   3%|▎         | 21/809 [00:24<15:38,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.282     0.3948     0.1298        136        640:   3%|▎         | 23/809 [00:27<15:19,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.275     0.3982       0.13        346        640:   3%|▎         | 25/809 [00:29<15:12,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.277     0.3986     0.1295        134        640:   3%|▎         | 26/809 [00:30<15:07,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.296     0.3955     0.1484        364        640:   3%|▎         | 28/809 [00:32<15:15,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.309      0.392     0.1524        177        640:   4%|▎         | 30/809 [00:35<15:27,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.317     0.3879      0.152        133        640:   4%|▍         | 31/809 [00:36<15:17,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.307     0.3911       0.15        237        640:   4%|▍         | 33/809 [00:38<15:02,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.304     0.3911     0.1478        294        640:   4%|▍         | 35/809 [00:41<14:52,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.316     0.3878     0.1514        547        640:   4%|▍         | 36/809 [00:42<15:02,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.294     0.3971     0.1479         93        640:   5%|▍         | 38/809 [00:44<14:59,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.271     0.4032     0.1452        156        640:   5%|▍         | 40/809 [00:46<14:47,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.255      0.408     0.1435        281        640:   5%|▌         | 42/809 [00:49<14:47,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.251     0.4086     0.1435        314        640:   5%|▌         | 43/809 [00:50<14:53,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.237     0.4124     0.1415        208        640:   6%|▌         | 45/809 [00:52<14:49,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.227     0.4156     0.1402        345        640:   6%|▌         | 47/809 [00:55<15:02,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.209     0.4217     0.1378        165        640:   6%|▌         | 49/809 [00:57<14:41,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.204     0.4227     0.1368        126        640:   6%|▌         | 50/809 [00:58<14:36,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.191     0.4255     0.1361        157        640:   6%|▋         | 52/809 [01:00<14:35,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.183     0.4282     0.1351        215        640:   7%|▋         | 54/809 [01:03<14:33,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.183     0.4278     0.1344        337        640:   7%|▋         | 56/809 [01:05<14:33,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.179     0.4293     0.1338        214        640:   7%|▋         | 57/809 [01:06<14:28,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.174      0.431     0.1329        113        640:   7%|▋         | 59/809 [01:08<14:24,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.173     0.4313     0.1319        225        640:   8%|▊         | 61/809 [01:11<14:46,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.171     0.4328     0.1318        212        640:   8%|▊         | 62/809 [01:12<14:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.169     0.4349     0.1313        175        640:   8%|▊         | 64/809 [01:14<14:39,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.189     0.4298     0.1385        440        640:   8%|▊         | 66/809 [01:17<15:00,  1.21s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.189     0.4299     0.1383        284        640:   8%|▊         | 67/809 [01:18<14:49,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.19     0.4291     0.1382        195        640:   9%|▊         | 69/809 [01:20<14:40,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.193     0.4284     0.1385        187        640:   9%|▉         | 71/809 [01:23<14:16,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G        1.2     0.4278      0.143        495        640:   9%|▉         | 73/809 [01:25<14:23,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.194     0.4293     0.1424        158        640:   9%|▉         | 74/809 [01:26<14:14,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.183      0.433     0.1411        300        640:   9%|▉         | 76/809 [01:28<14:14,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.184     0.4328     0.1416        130        640:  10%|▉         | 78/809 [01:31<14:15,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.199     0.4288     0.1454        127        640:  10%|▉         | 80/809 [01:33<14:00,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.204     0.4275     0.1466        109        640:  10%|█         | 81/809 [01:34<13:56,  1.15s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.22     0.4228     0.1486        291        640:  10%|█         | 83/809 [01:37<14:02,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.235      0.419      0.152        275        640:  11%|█         | 85/809 [01:39<14:10,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.236     0.4194     0.1524        133        640:  11%|█         | 86/809 [01:40<13:59,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.24     0.4186      0.153         99        640:  11%|█         | 88/809 [01:42<13:54,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.251     0.4159     0.1549        246        640:  11%|█         | 90/809 [01:45<13:57,  1.16s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4119     0.1559        241        640:  11%|█▏        | 92/809 [01:47<14:08,  1.18s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.269     0.4102     0.1563        220        640:  11%|█▏        | 93/809 [01:48<14:14,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4124     0.1554         96        640:  12%|█▏        | 95/809 [01:51<13:58,  1.17s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.266     0.4117     0.1577         72        640:  12%|█▏        | 97/809 [01:53<14:14,  1.20s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4125     0.1572        221        640:  12%|█▏        | 98/809 [01:54<14:04,  1.19s/it]d_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.268      0.411     0.1578        468        640:  12%|█▏        | 100/809 [01:57<13:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.269     0.4104     0.1584        241        640:  13%|█▎        | 102/809 [01:59<14:29,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.268     0.4106     0.1583        216        640:  13%|█▎        | 103/809 [02:00<14:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.27      0.411     0.1611         79        640:  13%|█▎        | 105/809 [02:03<14:13,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4128     0.1602         77        640:  13%|█▎        | 107/809 [02:05<13:56,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.266     0.4116     0.1604        386        640:  13%|█▎        | 109/809 [02:07<13:58,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.266     0.4114     0.1601        301        640:  14%|█▎        | 110/809 [02:09<13:42,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.266     0.4111     0.1599        179        640:  14%|█▍        | 112/809 [02:11<13:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.271     0.4092     0.1601        407        640:  14%|█▍        | 114/809 [02:13<13:38,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.274     0.4076     0.1601        225        640:  14%|█▍        | 116/809 [02:16<13:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.28     0.4061      0.162        545        640:  14%|█▍        | 117/809 [02:17<13:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.274     0.4076     0.1612        278        640:  15%|█▍        | 119/809 [02:19<13:35,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.271     0.4085     0.1607        221        640:  15%|█▍        | 120/809 [02:21<13:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.276     0.4071     0.1645        166        640:  15%|█▌        | 122/809 [02:23<13:44,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.276     0.4063     0.1645        431        640:  15%|█▌        | 124/809 [02:25<13:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.273     0.4069     0.1635        197        640:  16%|█▌        | 126/809 [02:28<13:19,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.273     0.4069     0.1636        182        640:  16%|█▌        | 127/809 [02:29<13:17,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.276     0.4052     0.1636        292        640:  16%|█▌        | 129/809 [02:31<13:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.275     0.4054      0.163        325        640:  16%|█▌        | 131/809 [02:33<13:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.277     0.4048      0.163        333        640:  16%|█▋        | 132/809 [02:35<13:15,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.279      0.404     0.1629        249        640:  17%|█▋        | 134/809 [02:37<13:31,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.275     0.4048      0.162        167        640:  17%|█▋        | 136/809 [02:39<13:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.272     0.4055     0.1614        480        640:  17%|█▋        | 138/809 [02:42<13:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.27     0.4057     0.1609        189        640:  17%|█▋        | 139/809 [02:43<13:03,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.268     0.4064     0.1604        217        640:  17%|█▋        | 141/809 [02:45<13:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.27     0.4057     0.1608         89        640:  18%|█▊        | 143/809 [02:48<12:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.268     0.4061     0.1607        246        640:  18%|█▊        | 145/809 [02:50<12:41,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.27     0.4055     0.1606        345        640:  18%|█▊        | 146/809 [02:51<12:41,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.27     0.4054     0.1603        119        640:  18%|█▊        | 148/809 [02:53<12:35,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.269     0.4055     0.1599        342        640:  19%|█▊        | 150/809 [02:56<12:38,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.272     0.4047     0.1609        361        640:  19%|█▊        | 151/809 [02:57<12:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.27     0.4052     0.1606        191        640:  19%|█▉        | 153/809 [02:59<12:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.271     0.4047     0.1614        256        640:  19%|█▉        | 155/809 [03:02<13:03,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.27     0.4056      0.161        147        640:  19%|█▉        | 156/809 [03:03<12:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.272     0.4046     0.1615        464        640:  20%|█▉        | 158/809 [03:05<12:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.272     0.4044     0.1616        358        640:  20%|█▉        | 160/809 [03:07<12:38,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.271     0.4046     0.1614        224        640:  20%|██        | 162/809 [03:10<12:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.271     0.4042     0.1612        227        640:  20%|██        | 163/809 [03:11<12:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.273     0.4035     0.1621        281        640:  20%|██        | 165/809 [03:13<12:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.271      0.404     0.1616        265        640:  21%|██        | 167/809 [03:16<12:44,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.269     0.4043     0.1613        170        640:  21%|██        | 168/809 [03:17<12:37,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.27     0.4035     0.1608        190        640:  21%|██        | 170/809 [03:19<12:28,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.269     0.4037     0.1602        126        640:  21%|██▏       | 172/809 [03:22<12:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.267     0.4038     0.1598        222        640:  21%|██▏       | 173/809 [03:23<12:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.265     0.4048     0.1591        173        640:  22%|██▏       | 175/809 [03:25<12:44,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.259     0.4067     0.1583        109        640:  22%|██▏       | 177/809 [03:28<12:22,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.26     0.4063      0.158        208        640:  22%|██▏       | 179/809 [03:30<12:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.26     0.4063      0.158        407        640:  22%|██▏       | 180/809 [03:31<12:12,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.261     0.4058     0.1578        280        640:  22%|██▏       | 182/809 [03:33<12:30,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4053     0.1582        613        640:  23%|██▎       | 184/809 [03:36<12:12,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.265     0.4053     0.1582        179        640:  23%|██▎       | 185/809 [03:37<12:04,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.269     0.4045     0.1587        275        640:  23%|██▎       | 187/809 [03:39<12:02,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.269     0.4045     0.1587        125        640:  23%|██▎       | 189/809 [03:42<12:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.269     0.4043     0.1583        150        640:  24%|██▎       | 191/809 [03:44<11:57,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.269     0.4044     0.1581        163        640:  24%|██▎       | 192/809 [03:45<11:52,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.27     0.4042     0.1578        221        640:  24%|██▍       | 194/809 [03:47<11:52,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.268     0.4044     0.1576         99        640:  24%|██▍       | 196/809 [03:50<11:56,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.268     0.4046     0.1575        205        640:  24%|██▍       | 197/809 [03:51<12:03,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.267     0.4052     0.1571        276        640:  25%|██▍       | 199/809 [03:53<11:59,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.265     0.4055     0.1567        147        640:  25%|██▍       | 201/809 [03:56<11:51,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4062     0.1563        268        640:  25%|██▌       | 203/809 [03:58<11:52,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4061     0.1562        279        640:  25%|██▌       | 204/809 [03:59<11:51,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4071      0.156        167        640:  25%|██▌       | 206/809 [04:02<11:49,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4074     0.1558        254        640:  26%|██▌       | 208/809 [04:04<11:45,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4074     0.1557        138        640:  26%|██▌       | 209/809 [04:05<11:52,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.262     0.4077     0.1556        162        640:  26%|██▌       | 211/809 [04:07<11:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4074     0.1563        476        640:  26%|██▋       | 213/809 [04:10<12:04,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.262      0.408     0.1561        225        640:  26%|██▋       | 214/809 [04:11<11:48,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4077      0.156        316        640:  27%|██▋       | 216/809 [04:13<11:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264      0.408     0.1559        233        640:  27%|██▋       | 218/809 [04:16<11:41,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4081     0.1557        148        640:  27%|██▋       | 220/809 [04:18<11:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4088     0.1554        111        640:  27%|██▋       | 221/809 [04:19<11:34,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4086     0.1554        271        640:  28%|██▊       | 223/809 [04:22<11:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4088     0.1552        229        640:  28%|██▊       | 225/809 [04:24<11:14,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4089     0.1551        131        640:  28%|██▊       | 226/809 [04:25<11:11,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4095     0.1546        249        640:  28%|██▊       | 228/809 [04:27<11:07,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4095     0.1543        237        640:  28%|██▊       | 230/809 [04:30<11:08,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4101     0.1541        236        640:  29%|██▊       | 232/809 [04:32<11:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4103     0.1539        228        640:  29%|██▉       | 233/809 [04:33<11:08,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.262     0.4108     0.1535        179        640:  29%|██▉       | 235/809 [04:35<11:10,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.261     0.4115      0.153         74        640:  29%|██▉       | 237/809 [04:38<11:28,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.262     0.4111     0.1529        249        640:  29%|██▉       | 238/809 [04:39<11:21,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.262     0.4109     0.1526        326        640:  30%|██▉       | 240/809 [04:41<11:14,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4104     0.1524        437        640:  30%|██▉       | 242/809 [04:44<11:08,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.264     0.4106     0.1521        262        640:  30%|███       | 244/809 [04:46<11:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.265     0.4102     0.1521        164        640:  30%|███       | 245/809 [04:47<11:22,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4112     0.1517        206        640:  31%|███       | 247/809 [04:50<11:47,  1.26s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.263     0.4114     0.1518        212        640:  31%|███       | 248/809 [04:51<11:28,  1.23s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.261     0.4121     0.1514        199        640:  31%|███       | 250/809 [04:53<11:07,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.259      0.413     0.1511        289        640:  31%|███       | 252/809 [04:56<10:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.257     0.4137     0.1507        283        640:  31%|███▏      | 254/809 [04:58<10:49,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.255     0.4139     0.1503        209        640:  32%|███▏      | 256/809 [05:00<10:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.255     0.4139     0.1501        334        640:  32%|███▏      | 257/809 [05:02<10:41,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.255      0.414     0.1502         90        640:  32%|███▏      | 259/809 [05:04<11:05,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.253     0.4142     0.1498        183        640:  32%|███▏      | 261/809 [05:06<10:51,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.253     0.4146     0.1496        151        640:  32%|███▏      | 262/809 [05:08<10:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.252     0.4151     0.1493        183        640:  33%|███▎      | 264/809 [05:10<10:46,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.252      0.415     0.1492        259        640:  33%|███▎      | 266/809 [05:12<10:31,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.254     0.4145     0.1494        472        640:  33%|███▎      | 268/809 [05:15<10:27,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.254     0.4145     0.1494        211        640:  33%|███▎      | 269/809 [05:16<10:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.254     0.4148      0.149        241        640:  33%|███▎      | 271/809 [05:18<10:26,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.255     0.4145     0.1498        183        640:  34%|███▎      | 273/809 [05:20<10:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.254     0.4149     0.1497        113        640:  34%|███▍      | 275/809 [05:23<10:17,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.253     0.4152     0.1496        198        640:  34%|███▍      | 276/809 [05:24<10:16,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.252     0.4159     0.1493        202        640:  34%|███▍      | 278/809 [05:26<10:40,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.252     0.4161     0.1494        430        640:  35%|███▍      | 280/809 [05:29<10:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.252      0.416     0.1493        371        640:  35%|███▍      | 281/809 [05:30<10:28,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.252     0.4161     0.1491        367        640:  35%|███▍      | 283/809 [05:32<10:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.249      0.417      0.149         92        640:  35%|███▌      | 285/809 [05:35<10:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.25     0.4169     0.1492        360        640:  35%|███▌      | 286/809 [05:36<10:25,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.249     0.4172     0.1489        274        640:  36%|███▌      | 288/809 [05:38<10:14,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.248      0.418     0.1487        260        640:  36%|███▌      | 290/809 [05:41<10:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.248     0.4179     0.1486        226        640:  36%|███▌      | 291/809 [05:42<10:11,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.248     0.4183     0.1483        237        640:  36%|███▌      | 293/809 [05:44<10:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.247     0.4185      0.148        269        640:  36%|███▋      | 295/809 [05:47<10:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.246     0.4187     0.1478        232        640:  37%|███▋      | 296/809 [05:48<10:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.245     0.4191     0.1474        310        640:  37%|███▋      | 298/809 [05:50<09:59,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.243     0.4195     0.1471        223        640:  37%|███▋      | 300/809 [05:52<09:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.241     0.4205     0.1467        198        640:  37%|███▋      | 302/809 [05:55<09:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.24     0.4206     0.1466        237        640:  37%|███▋      | 303/809 [05:56<09:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.239     0.4209     0.1464        304        640:  38%|███▊      | 305/809 [05:58<09:53,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.239     0.4211     0.1461        256        640:  38%|███▊      | 307/809 [06:01<09:44,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.238     0.4212     0.1459        132        640:  38%|███▊      | 309/809 [06:03<09:46,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.238     0.4211     0.1463        547        640:  38%|███▊      | 310/809 [06:04<09:54,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.236     0.4219      0.146        137        640:  39%|███▊      | 312/809 [06:06<09:47,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.234     0.4224     0.1457        158        640:  39%|███▉      | 314/809 [06:09<09:46,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.234     0.4225     0.1456        319        640:  39%|███▉      | 315/809 [06:10<09:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4232     0.1452        219        640:  39%|███▉      | 317/809 [06:12<09:49,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.23     0.4237     0.1449        286        640:  39%|███▉      | 319/809 [06:15<09:40,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.23     0.4236     0.1447        339        640:  40%|███▉      | 320/809 [06:16<09:51,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4235     0.1449        344        640:  40%|███▉      | 322/809 [06:18<09:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4233     0.1447        318        640:  40%|████      | 324/809 [06:21<09:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4229     0.1453        381        640:  40%|████      | 325/809 [06:22<09:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4232     0.1452        204        640:  40%|████      | 327/809 [06:24<09:25,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4234     0.1449        166        640:  41%|████      | 329/809 [06:27<09:30,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4232     0.1452        523        640:  41%|████      | 331/809 [06:29<09:21,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4233     0.1451        300        640:  41%|████      | 332/809 [06:30<09:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.233      0.423     0.1451        276        640:  41%|████▏     | 334/809 [06:32<09:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.233     0.4228     0.1449        153        640:  42%|████▏     | 336/809 [06:35<09:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.234     0.4224     0.1449        389        640:  42%|████▏     | 337/809 [06:36<09:30,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.235      0.422     0.1449         85        640:  42%|████▏     | 339/809 [06:38<09:16,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.235     0.4219     0.1447        217        640:  42%|████▏     | 341/809 [06:41<09:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.235     0.4218     0.1445        231        640:  42%|████▏     | 342/809 [06:42<09:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.236     0.4216     0.1447        346        640:  43%|████▎     | 344/809 [06:44<09:11,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.237     0.4213     0.1449        346        640:  43%|████▎     | 346/809 [06:47<09:05,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.237     0.4211     0.1449        283        640:  43%|████▎     | 347/809 [06:48<09:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.238      0.421     0.1449        392        640:  43%|████▎     | 349/809 [06:50<09:08,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.236     0.4212     0.1446        271        640:  43%|████▎     | 351/809 [06:53<09:06,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.235     0.4213     0.1444        319        640:  44%|████▎     | 353/809 [06:55<08:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.235     0.4212     0.1442        272        640:  44%|████▍     | 354/809 [06:56<08:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.234     0.4215     0.1439        207        640:  44%|████▍     | 356/809 [06:59<08:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.234     0.4213     0.1439        434        640:  44%|████▍     | 358/809 [07:01<08:55,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.234     0.4213     0.1439        242        640:  44%|████▍     | 359/809 [07:02<08:48,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.234     0.4215     0.1437        333        640:  45%|████▍     | 361/809 [07:04<08:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4219     0.1434         73        640:  45%|████▍     | 363/809 [07:07<08:44,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4223     0.1433        190        640:  45%|████▍     | 364/809 [07:08<08:41,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4221     0.1435        348        640:  45%|████▌     | 366/809 [07:10<08:54,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.233     0.4217     0.1435        321        640:  45%|████▌     | 368/809 [07:13<08:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4218     0.1434        231        640:  46%|████▌     | 369/809 [07:14<08:34,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4219     0.1432        380        640:  46%|████▌     | 371/809 [07:16<08:36,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.233     0.4215     0.1432        534        640:  46%|████▌     | 373/809 [07:19<08:36,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.234     0.4211     0.1436        222        640:  46%|████▋     | 375/809 [07:21<08:37,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.233     0.4217     0.1434         55        640:  46%|████▋     | 376/809 [07:22<08:27,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4219     0.1433        393        640:  47%|██���█▋     | 378/809 [07:25<08:20,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4221     0.1432        299        640:  47%|████▋     | 380/809 [07:27<08:26,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4222     0.1431        234        640:  47%|████▋     | 381/809 [07:28<08:32,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4221      0.143        223        640:  47%|████▋     | 383/809 [07:30<08:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4219     0.1429        262        640:  48%|████▊     | 385/809 [07:33<08:16,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231      0.422     0.1428        177        640:  48%|████▊     | 387/809 [07:35<08:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4217     0.1432        358        640:  48%|████▊     | 388/809 [07:36<08:09,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4219      0.143        102        640:  48%|████▊     | 390/809 [07:39<08:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.23     0.4222     0.1428        214        640:  48%|████▊     | 392/809 [07:41<08:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.23     0.4222     0.1427        291        640:  49%|████▊     | 393/809 [07:42<08:03,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.228     0.4226     0.1424        308        640:  49%|████▉     | 395/809 [07:45<08:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.227     0.4229     0.1423         86        640:  49%|████▉     | 397/809 [07:47<08:22,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.227     0.4226     0.1423        237        640:  49%|████▉     | 398/809 [07:48<08:17,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.228     0.4225     0.1423        377        640:  49%|████▉     | 400/809 [07:51<08:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.229      0.422     0.1424        273        640:  50%|████▉     | 402/809 [07:53<08:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.229     0.4218     0.1422        212        640:  50%|████▉     | 404/809 [07:55<07:52,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.228     0.4219     0.1422        113        640:  50%|█████     | 406/809 [07:58<07:49,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.228     0.4218     0.1421        328        640:  50%|█████     | 407/809 [07:59<07:46,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.229     0.4215     0.1422        260        640:  51%|█████     | 409/809 [08:01<07:50,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4209     0.1423        281        640:  51%|█████     | 411/809 [08:04<07:40,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231      0.421     0.1423        121        640:  51%|█████     | 412/809 [08:05<07:36,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4207     0.1424        181        640:  51%|█████     | 414/809 [08:07<07:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4209     0.1423        238        640:  51%|█████▏    | 416/809 [08:09<07:39,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.232     0.4204     0.1425        268        640:  52%|█████▏    | 418/809 [08:12<07:39,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.233     0.4202     0.1425        279        640:  52%|█████▏    | 419/809 [08:13<07:33,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.234     0.4198     0.1431        399        640:  52%|█████▏    | 421/809 [08:15<07:40,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.235     0.4195     0.1431        244        640:  52%|█████▏    | 423/809 [08:18<07:33,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.235     0.4193     0.1431        271        640:  52%|█████▏    | 424/809 [08:19<07:38,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.236      0.419     0.1432        518        640:  53%|█████▎    | 426/809 [08:21<07:35,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.237     0.4186     0.1433        367        640:  53%|█████▎    | 428/809 [08:23<07:22,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.237     0.4186     0.1433        197        640:  53%|█████▎    | 430/809 [08:26<07:17,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.237     0.4186     0.1435        310        640:  53%|█████▎    | 431/809 [08:27<07:15,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.235      0.419     0.1433        172        640:  54%|█████▎    | 433/809 [08:29<07:13,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.234     0.4191     0.1431        191        640:  54%|█████▍    | 435/809 [08:32<07:10,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.235     0.4189     0.1445        506        640:  54%|█████▍    | 436/809 [08:33<07:20,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.235      0.419     0.1444        142        640:  54%|█████▍    | 438/809 [08:35<07:23,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.233     0.4194     0.1444        206        640:  54%|█████▍    | 440/809 [08:37<07:11,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.233     0.4195     0.1443        296        640:  55%|█████▍    | 441/809 [08:39<07:06,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.234     0.4191     0.1449        283        640:  55%|█████▍    | 443/809 [08:41<07:13,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.233     0.4192     0.1449        180        640:  55%|█████▌    | 445/809 [08:43<07:12,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4198     0.1447        193        640:  55%|█████▌    | 447/809 [08:46<07:09,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4199     0.1447        252        640:  55%|█████▌    | 448/809 [08:47<07:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.231     0.4198     0.1448        525        640:  56%|█████▌    | 450/809 [08:49<07:09,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.23     0.4198     0.1447        333        640:  56%|█████▌    | 452/809 [08:52<07:16,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.229     0.4201     0.1445        158        640:  56%|█████▌    | 453/809 [08:53<07:14,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.23     0.4199     0.1449        244        640:  56%|█████▌    | 455/809 [08:55<07:02,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.229       0.42     0.1451        291        640:  56%|█████▋    | 457/809 [08:58<06:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.228     0.4203     0.1449        134        640:  57%|█████▋    | 458/809 [08:59<06:48,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.228     0.4205     0.1447        249        640:  57%|█████▋    | 460/809 [09:01<06:42,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.227     0.4208     0.1445        417        640:  57%|█████▋    | 462/809 [09:03<06:40,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.227     0.4209     0.1446        389        640:  57%|█████▋    | 464/809 [09:06<06:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.226      0.421     0.1445        274        640:  57%|█████▋    | 465/809 [09:07<06:37,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.225     0.4213     0.1443        218        640:  58%|█████▊    | 467/809 [09:09<06:35,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.224     0.4216     0.1441         87        640:  58%|█████▊    | 469/809 [09:12<06:34,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.224     0.4218     0.1438        182        640:  58%|█████▊    | 471/809 [09:14<06:28,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.224     0.4216      0.144        505        640:  58%|█████▊    | 472/809 [09:15<06:27,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.224     0.4218     0.1439        397        640:  59%|█████▊    | 474/809 [09:17<06:25,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.223      0.422     0.1438        175        640:  59%|█████▉    | 476/809 [09:20<06:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.224     0.4217     0.1438        288        640:  59%|█████▉    | 478/809 [09:22<06:24,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.225     0.4216     0.1439        365        640:  59%|█████▉    | 479/809 [09:23<06:33,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.223     0.4221     0.1436        249        640:  59%|█████▉    | 481/809 [09:26<06:28,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.222     0.4223     0.1434        347        640:  60%|█████▉    | 483/809 [09:28<06:29,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.222     0.4225     0.1433        135        640:  60%|█████▉    | 484/809 [09:29<06:37,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.222     0.4225     0.1432        349        640:  60%|██████    | 486/809 [09:32<06:24,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.223     0.4223     0.1439        473        640:  60%|██████    | 487/809 [09:33<06:26,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.223     0.4223     0.1438        256        640:  60%|██████    | 489/809 [09:35<06:15,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.225      0.422     0.1443        754        640:  61%|██████    | 491/809 [09:37<06:17,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.224     0.4222     0.1443        318        640:  61%|██████    | 493/809 [09:40<06:20,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.224     0.4224     0.1441        188        640:  61%|██████    | 494/809 [09:41<06:13,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.223     0.4226     0.1439        249        640:  61%|██████▏   | 496/809 [09:43<06:07,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.222     0.4228     0.1437        171        640:  62%|██████▏   | 498/809 [09:46<06:04,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.223     0.4226      0.144        496        640:  62%|██████▏   | 499/809 [09:47<06:11,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.223     0.4225     0.1441        364        640:  62%|██████▏   | 501/809 [09:49<06:11,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.222     0.4231     0.1438        196        640:  62%|██████▏   | 503/809 [09:52<06:01,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.22     0.4238     0.1436        171        640:  62%|██████▏   | 505/809 [09:54<05:53,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G       1.22     0.4239     0.1435        274        640:  63%|██████▎   | 506/809 [09:55<05:57,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.219     0.4242     0.1433        250        640:  63%|██████▎   | 508/809 [09:58<05:54,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.218     0.4247     0.1433        168        640:  63%|██████▎   | 510/809 [10:00<05:50,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.219     0.4246     0.1435        307        640:  63%|██████▎   | 511/809 [10:01<05:59,  1.21s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.219     0.4245     0.1436        556        640:  63%|██████▎   | 513/809 [10:04<05:55,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.219     0.4246     0.1435        290        640:  64%|██████▎   | 515/809 [10:06<05:43,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.219     0.4247     0.1434        276        640:  64%|██████▍   | 516/809 [10:07<05:38,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.217     0.4251     0.1432        183        640:  64%|██████▍   | 518/809 [10:09<05:40,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.218      0.425     0.1433        177        640:  64%|██████▍   | 520/809 [10:12<05:32,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.218     0.4252     0.1432        140        640:  65%|██████▍   | 522/809 [10:14<05:30,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.217     0.4255     0.1431        208        640:  65%|██████▍   | 523/809 [10:15<05:29,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.216     0.4261     0.1429         99        640:  65%|██████▍   | 525/809 [10:18<05:41,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.216     0.4263     0.1428        345        640:  65%|██████▌   | 527/809 [10:20<05:34,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.215     0.4264     0.1427        368        640:  65%|██████▌   | 528/809 [10:21<05:29,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.215     0.4266     0.1427        414        640:  66%|██████▌   | 530/809 [10:23<05:21,  1.15s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.215     0.4268     0.1425        373        640:  66%|██████▌   | 532/809 [10:26<05:16,  1.14s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.214     0.4271     0.1423        174        640:  66%|██████▌   | 534/809 [10:28<05:21,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.214     0.4273     0.1422        100        640:  66%|██████▌   | 535/809 [10:29<05:24,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.213     0.4275     0.1421        272        640:  66%|██████▋   | 537/809 [10:32<05:31,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.213     0.4276     0.1419        213        640:  67%|██████▋   | 539/809 [10:34<05:19,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.213     0.4275      0.142        449        640:  67%|██████▋   | 540/809 [10:35<05:18,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.214     0.4274      0.142        267        640:  67%|██████▋   | 542/809 [10:38<05:24,  1.22s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.214     0.4276     0.1418        155        640:  67%|██████▋   | 544/809 [10:40<05:18,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.213      0.428     0.1417         75        640:  67%|██████▋   | 545/809 [10:41<05:12,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.212     0.4286     0.1415        165        640:  68%|██████▊   | 547/809 [10:44<05:07,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.212     0.4284     0.1416        271        640:  68%|██████▊   | 549/809 [10:46<05:06,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.213     0.4282     0.1416        365        640:  68%|██████▊   | 550/809 [10:47<05:02,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.213     0.4281     0.1414        214        640:  68%|██████▊   | 552/809 [10:49<05:01,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.213      0.428     0.1413        160        640:  68%|██████▊   | 554/809 [10:52<05:00,  1.18s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.214     0.4279     0.1412        159        640:  69%|██████▊   | 556/809 [10:54<04:54,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.213     0.4282      0.141        236        640:  69%|██████▉   | 558/809 [10:56<04:51,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.213     0.4282      0.141        420        640:  69%|██████▉   | 559/809 [10:58<04:50,  1.16s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.212     0.4288     0.1408        242        640:  69%|██████▉   | 561/809 [11:00<04:57,  1.20s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.211      0.429     0.1406        263        640:  70%|██████▉   | 563/809 [11:02<04:47,  1.17s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)
-     89/300      13.6G      1.211     0.4293     0.1405        227        640:  70%|██████▉   | 565/809 [11:05<04:50,  1.19s/it]_sampler_2d_backward_cuda does not have a deterministic implementation, but you set 'torch.use_deterministic_algorithms(True, warn_only=True)'. You can file an issue at https://github.com/pytorch/pytorch/issues to help us prioritize adding deterministic support for this operation. (Triggered internally at /usr/local/src/pytorch/aten/src/ATen/Context.cpp:71.)