diff --git "a/wandb/run-20220504_142129-1tmxz74i/files/output.log" "b/wandb/run-20220504_142129-1tmxz74i/files/output.log"
new file mode 100644--- /dev/null
+++ "b/wandb/run-20220504_142129-1tmxz74i/files/output.log"
@@ -0,0 +1,5053 @@
+wandb: WARNING Config item 'output_dir' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'evaluation_strategy' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'per_device_train_batch_size' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'per_device_eval_batch_size' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'gradient_accumulation_steps' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'learning_rate' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'num_train_epochs' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'warmup_steps' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'logging_steps' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'save_steps' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'eval_steps' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'metric_for_best_model' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'greater_is_better' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'generation_max_length' was locked by 'sweep' (ignored update).
+wandb: WARNING Config item 'generation_num_beams' was locked by 'sweep' (ignored update).
+  0%|                                                                                          | 0/9720 [00:00<?, ?it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 11.5079, 'learning_rate': 0.0, 'epoch': 0.0}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                               | 1/9720 [00:08<24:15:50,  8.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 11.1203, 'learning_rate': 0.0, 'epoch': 0.0}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                               | 2/9720 [00:16<21:35:11,  8.00s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 10.6958, 'learning_rate': 0.0, 'epoch': 0.0}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                               | 3/9720 [00:23<20:11:02,  7.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 11.2693, 'learning_rate': 0.0, 'epoch': 0.0}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                               | 4/9720 [00:29<19:15:21,  7.13s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                               | 5/9720 [00:36<18:29:29,  6.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 11.6199, 'learning_rate': 0.0, 'epoch': 0.0}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                               | 6/9720 [00:42<17:56:22,  6.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 10.6054, 'learning_rate': 6.569998523345044e-07, 'epoch': 0.0}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                               | 7/9720 [00:48<17:26:14,  6.46s/it]
+  0%|                                                                               | 7/9720 [00:48<17:26:14,  6.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 10.9872, 'learning_rate': 1.3139997046690089e-06, 'epoch': 0.0}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                               | 8/9720 [00:54<16:52:30,  6.26s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 10.8356, 'learning_rate': 1.9709995570035133e-06, 'epoch': 0.0}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                               | 9/9720 [00:59<16:24:14,  6.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 11.1359, 'learning_rate': 2.6279994093380178e-06, 'epoch': 0.0}
+  0%|                                                                              | 10/9720 [01:05<15:59:23,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                              | 11/9720 [01:10<15:35:50,  5.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 10.0616, 'learning_rate': 3.284999261672522e-06, 'epoch': 0.0}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 9.5432, 'learning_rate': 3.941999114007027e-06, 'epoch': 0.0}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                              | 12/9720 [01:16<15:11:42,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 9.1687, 'learning_rate': 4.598998966341531e-06, 'epoch': 0.0}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                              | 13/9720 [01:22<15:16:40,  5.67s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                              | 14/9720 [01:27<14:57:52,  5.55s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 9.1152, 'learning_rate': 5.2559988186760355e-06, 'epoch': 0.0}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 8.438, 'learning_rate': 5.9129986710105395e-06, 'epoch': 0.0}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|                                                                              | 15/9720 [01:32<14:37:09,  5.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 16/9720 [01:37<14:26:59,  5.36s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 8.3793, 'learning_rate': 6.569998523345044e-06, 'epoch': 0.0}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 7.4295, 'learning_rate': 7.2269983756795484e-06, 'epoch': 0.01}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 17/9720 [01:42<14:07:02,  5.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 18/9720 [01:47<13:47:12,  5.12s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 7.3944, 'learning_rate': 7.883998228014053e-06, 'epoch': 0.01}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 19/9720 [01:52<13:30:59,  5.02s/it]
+  0%|▏                                                                             | 19/9720 [01:52<13:30:59,  5.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 20/9720 [01:56<13:12:10,  4.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 6.8123, 'learning_rate': 9.197997932683061e-06, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 21/9720 [02:01<12:59:44,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 6.2127, 'learning_rate': 9.854997785017565e-06, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 22/9720 [02:06<12:45:20,  4.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 6.5409, 'learning_rate': 1.0511997637352071e-05, 'epoch': 0.01}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.9446, 'learning_rate': 1.1168997489686577e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 23/9720 [02:10<12:30:35,  4.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 5.9737, 'learning_rate': 1.1825997342021079e-05, 'epoch': 0.01}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 24/9720 [02:14<12:16:35,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 25/9720 [02:19<12:25:46,  4.62s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.7396, 'learning_rate': 1.2482997194355583e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 26/9720 [02:23<12:13:28,  4.54s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 5.4264, 'learning_rate': 1.3139997046690089e-05, 'epoch': 0.01}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 27/9720 [02:28<12:00:10,  4.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 5.2598, 'learning_rate': 1.3796996899024595e-05, 'epoch': 0.01}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 28/9720 [02:32<11:45:19,  4.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.4661, 'learning_rate': 1.4453996751359097e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 29/9720 [02:36<11:34:15,  4.30s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.965, 'learning_rate': 1.5110996603693601e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 30/9720 [02:40<11:23:03,  4.23s/it]
+  0%|▏                                                                             | 30/9720 [02:40<11:23:03,  4.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▏                                                                             | 31/9720 [02:44<11:11:07,  4.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.2457, 'learning_rate': 1.642499630836261e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                             | 32/9720 [02:48<11:00:32,  4.09s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.7293, 'learning_rate': 1.7081996160697115e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                             | 33/9720 [02:52<10:49:59,  4.03s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.7455, 'learning_rate': 1.773899601303162e-05, 'epoch': 0.01}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                             | 34/9720 [02:56<10:40:47,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.8807, 'learning_rate': 1.8395995865366123e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                             | 35/9720 [02:59<10:32:22,  3.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 5.0133, 'learning_rate': 1.905299571770063e-05, 'epoch': 0.01}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                             | 36/9720 [03:03<10:23:15,  3.86s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5158, 'learning_rate': 1.970999557003513e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.6027, 'learning_rate': 2.0366995422369638e-05, 'epoch': 0.01}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                             | 37/9720 [03:07<10:12:43,  3.80s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.5853, 'learning_rate': 2.1023995274704142e-05, 'epoch': 0.01}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                             | 38/9720 [03:11<10:26:03,  3.88s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                             | 39/9720 [03:15<10:17:10,  3.83s/it]
+  0%|▎                                                                             | 39/9720 [03:15<10:17:10,  3.83s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [03:15<10:17:10,  3.83s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                             | 40/9720 [03:18<10:05:04,  3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [03:15<10:17:10,  3.83s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.432, 'learning_rate': 2.2337994979373154e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [03:15<10:17:10,  3.83s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                              | 41/9720 [03:22<9:51:44,  3.67s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [03:15<10:17:10,  3.83s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1745, 'learning_rate': 2.2994994831707658e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [03:15<10:17:10,  3.83s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3357, 'learning_rate': 2.3651994684042158e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [03:15<10:17:10,  3.83s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                              | 43/9720 [03:28<9:28:08,  3.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [03:15<10:17:10,  3.83s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1965, 'learning_rate': 2.4308994536376662e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  0%|▎                                                                              | 44/9720 [03:32<9:16:14,  3.45s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  0%|▎                                                                              | 44/9720 [03:32<9:16:14,  3.45s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [03:32<9:16:14,  3.45s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [03:32<9:16:14,  3.45s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.0005, 'learning_rate': 2.5622994241045674e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [03:32<9:16:14,  3.45s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▎                                                                              | 46/9720 [03:38<8:49:35,  3.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [03:32<9:16:14,  3.45s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.9979, 'learning_rate': 2.6279994093380178e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [03:32<9:16:14,  3.45s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.808, 'learning_rate': 2.693699394571468e-05, 'epoch': 0.01}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [03:32<9:16:14,  3.45s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  0%|▍                                                                              | 48/9720 [03:44<8:20:43,  3.11s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [03:32<9:16:14,  3.45s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  1%|▍                                                                              | 49/9720 [03:47<8:05:09,  3.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                              | 49/9720 [03:47<8:05:09,  3.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.2427, 'learning_rate': 2.8250993650383686e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                              | 50/9720 [03:50<8:07:33,  3.03s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.2449, 'learning_rate': 2.8907993502718194e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                             | 51/9720 [03:58<12:05:41,  4.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 6.2136, 'learning_rate': 2.9564993355052698e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 5.8481, 'learning_rate': 3.0221993207387202e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                             | 52/9720 [04:05<14:07:58,  5.26s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 5.5996, 'learning_rate': 3.0878993059721706e-05, 'epoch': 0.02}
+  1%|▍                                                                             | 53/9720 [04:11<15:09:48,  5.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.5013, 'learning_rate': 3.153599291205621e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                             | 54/9720 [04:18<15:43:54,  5.86s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.3008, 'learning_rate': 3.219299276439072e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                             | 55/9720 [04:24<15:59:13,  5.95s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.3777, 'learning_rate': 3.284999261672522e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                             | 56/9720 [04:30<16:01:23,  5.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                             | 57/9720 [04:36<15:49:38,  5.90s/it]
+  1%|▍                                                                             | 57/9720 [04:36<15:49:38,  5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.8636, 'learning_rate': 3.416399232139423e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                             | 58/9720 [04:41<15:33:33,  5.80s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.8621, 'learning_rate': 3.482099217372874e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                             | 59/9720 [04:47<15:18:57,  5.71s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.7417, 'learning_rate': 3.547799202606324e-05, 'epoch': 0.02}
+  1%|▍                                                                             | 60/9720 [04:52<15:06:35,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.8393, 'learning_rate': 3.6134991878397745e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                             | 61/9720 [04:57<14:49:25,  5.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6815, 'learning_rate': 3.6791991730732245e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▍                                                                             | 62/9720 [05:03<14:35:34,  5.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.8777, 'learning_rate': 3.744899158306675e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 63/9720 [05:08<14:40:06,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.8993, 'learning_rate': 3.810599143540126e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 64/9720 [05:13<14:22:10,  5.36s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 65/9720 [05:18<14:01:28,  5.23s/it]
+  1%|▌                                                                             | 65/9720 [05:18<14:01:28,  5.23s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6317, 'learning_rate': 3.941999114007026e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 66/9720 [05:23<13:47:01,  5.14s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 67/9720 [05:28<13:33:09,  5.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5564, 'learning_rate': 4.007699099240477e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5359, 'learning_rate': 4.0733990844739276e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 68/9720 [05:33<13:16:42,  4.95s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6462, 'learning_rate': 4.139099069707378e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 69/9720 [05:37<13:02:56,  4.87s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 70/9720 [05:42<12:52:10,  4.80s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.5479, 'learning_rate': 4.2047990549408284e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 71/9720 [05:46<12:39:14,  4.72s/it]
+  1%|▌                                                                             | 71/9720 [05:46<12:39:14,  4.72s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.4873, 'learning_rate': 4.336199025407729e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 72/9720 [05:51<12:29:10,  4.66s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5333, 'learning_rate': 4.40189901064118e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 73/9720 [05:56<12:21:22,  4.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.4942, 'learning_rate': 4.467598995874631e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 74/9720 [06:00<12:10:26,  4.54s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5229, 'learning_rate': 4.533298981108081e-05, 'epoch': 0.02}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 75/9720 [06:05<12:25:05,  4.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.2522, 'learning_rate': 4.5989989663415315e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 76/9720 [06:09<12:09:00,  4.54s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.4598, 'learning_rate': 4.664698951574981e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▌                                                                             | 77/9720 [06:13<11:53:37,  4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.3976, 'learning_rate': 4.7303989368084316e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 78/9720 [06:17<11:39:36,  4.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.3418, 'learning_rate': 4.796098922041882e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 79/9720 [06:22<11:27:19,  4.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.5355, 'learning_rate': 4.8617989072753324e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 80/9720 [06:26<11:14:42,  4.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.1474, 'learning_rate': 4.927498892508783e-05, 'epoch': 0.02}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 81/9720 [06:30<11:04:57,  4.14s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3487, 'learning_rate': 4.993198877742233e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 82/9720 [06:33<10:52:24,  4.06s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.2385, 'learning_rate': 5.058898862975684e-05, 'epoch': 0.03}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 83/9720 [06:37<10:40:15,  3.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.9705, 'learning_rate': 5.124598848209135e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 84/9720 [06:41<10:31:37,  3.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 85/9720 [06:45<10:22:44,  3.88s/it]
+  1%|▋                                                                             | 85/9720 [06:45<10:22:44,  3.88s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 86/9720 [06:48<10:13:03,  3.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.3691, 'learning_rate': 5.2559988186760355e-05, 'epoch': 0.03}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 87/9720 [06:52<10:06:11,  3.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2919, 'learning_rate': 5.321698803909486e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 88/9720 [06:56<10:24:01,  3.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.1494, 'learning_rate': 5.387398789142936e-05, 'epoch': 0.03}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.0668, 'learning_rate': 5.453098774376387e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                             | 89/9720 [07:00<10:14:41,  3.83s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.1362, 'learning_rate': 5.518798759609838e-05, 'epoch': 0.03}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                              | 90/9720 [07:04<9:59:55,  3.74s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                              | 91/9720 [07:07<9:46:43,  3.66s/it]
+  1%|▋                                                                              | 91/9720 [07:07<9:46:43,  3.66s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▋                                                                              | 92/9720 [07:10<9:35:41,  3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.0575, 'learning_rate': 5.650198730076737e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.9414, 'learning_rate': 5.715898715310188e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                              | 93/9720 [07:14<9:23:57,  3.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                              | 94/9720 [07:17<9:15:25,  3.46s/it]
+  1%|▊                                                                              | 94/9720 [07:17<9:15:25,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                              | 95/9720 [07:20<9:01:51,  3.38s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.8777, 'learning_rate': 5.847298685777089e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.4894, 'learning_rate': 5.9129986710105395e-05, 'epoch': 0.03}
+  1%|▊                                                                              | 96/9720 [07:23<8:44:02,  3.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.2646, 'learning_rate': 5.97869865624399e-05, 'epoch': 0.03}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                              | 97/9720 [07:26<8:29:39,  3.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                              | 98/9720 [07:29<8:15:19,  3.09s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.1012, 'learning_rate': 6.0443986414774403e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.3252, 'learning_rate': 6.110098626710891e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                              | 99/9720 [07:32<8:02:24,  3.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                             | 100/9720 [07:35<8:02:40,  3.01s/it]
+  1%|▊                                                                             | 100/9720 [07:35<8:02:40,  3.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                            | 101/9720 [07:43<11:59:46,  4.49s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 6.1741, 'learning_rate': 6.241498597177793e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 5.7244, 'learning_rate': 6.307198582411243e-05, 'epoch': 0.03}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                            | 102/9720 [07:50<13:59:58,  5.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.2937, 'learning_rate': 6.372898567644693e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                            | 103/9720 [07:56<15:01:16,  5.62s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                            | 104/9720 [08:03<15:34:31,  5.83s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.9471, 'learning_rate': 6.438598552878144e-05, 'epoch': 0.03}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                            | 105/9720 [08:09<15:49:29,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.8998, 'learning_rate': 6.504298538111594e-05, 'epoch': 0.03}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.0121, 'learning_rate': 6.569998523345044e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                            | 106/9720 [08:15<15:51:10,  5.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                            | 107/9720 [08:21<15:49:40,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.7973, 'learning_rate': 6.635698508578496e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.7891, 'learning_rate': 6.701398493811944e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                            | 108/9720 [08:26<15:34:24,  5.83s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                            | 109/9720 [08:32<15:17:59,  5.73s/it]
+  1%|▊                                                                            | 109/9720 [08:32<15:17:59,  5.73s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▊                                                                            | 110/9720 [08:37<15:02:33,  5.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.8595, 'learning_rate': 6.832798464278846e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4903, 'learning_rate': 6.898498449512296e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 111/9720 [08:43<14:50:57,  5.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 112/9720 [08:48<14:35:38,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.6337, 'learning_rate': 6.964198434745747e-05, 'epoch': 0.03}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 113/9720 [08:54<14:41:13,  5.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4506, 'learning_rate': 7.029898419979197e-05, 'epoch': 0.03}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 114/9720 [08:59<14:23:29,  5.39s/it]
+  1%|▉                                                                            | 114/9720 [08:59<14:23:29,  5.39s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 115/9720 [09:04<14:05:31,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.5671, 'learning_rate': 7.161298390446099e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 116/9720 [09:09<13:48:12,  5.17s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4214, 'learning_rate': 7.226998375679549e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6995, 'learning_rate': 7.292698360912999e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 117/9720 [09:13<13:34:45,  5.09s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 118/9720 [09:18<13:21:49,  5.01s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4702, 'learning_rate': 7.358398346146449e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4675, 'learning_rate': 7.4240983313799e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 119/9720 [09:23<13:06:24,  4.91s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5423, 'learning_rate': 7.48979831661335e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 120/9720 [09:28<12:54:47,  4.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 121/9720 [09:32<12:42:29,  4.77s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3751, 'learning_rate': 7.5554983018468e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 122/9720 [09:37<12:30:33,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3965, 'learning_rate': 7.621198287080252e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 123/9720 [09:41<12:18:05,  4.61s/it]
+  1%|▉                                                                            | 123/9720 [09:41<12:18:05,  4.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2151, 'learning_rate': 7.752598257547152e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 124/9720 [09:46<12:09:10,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 125/9720 [09:50<12:17:57,  4.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.2806, 'learning_rate': 7.818298242780602e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|▉                                                                            | 126/9720 [09:55<12:04:10,  4.53s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2132, 'learning_rate': 7.883998228014052e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.2403, 'learning_rate': 7.949698213247504e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                            | 127/9720 [09:59<11:49:41,  4.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.3122, 'learning_rate': 8.015398198480954e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                            | 128/9720 [10:03<11:33:47,  4.34s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.4731, 'learning_rate': 8.081098183714404e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                            | 129/9720 [10:07<11:19:54,  4.25s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.2146, 'learning_rate': 8.146798168947855e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                            | 130/9720 [10:11<11:07:33,  4.18s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.0394, 'learning_rate': 8.212498154181305e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                            | 131/9720 [10:15<11:04:43,  4.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1229, 'learning_rate': 8.278198139414755e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                            | 132/9720 [10:19<10:50:27,  4.07s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.9938, 'learning_rate': 8.343898124648207e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                            | 133/9720 [10:23<10:37:53,  3.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.0056, 'learning_rate': 8.409598109881657e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                            | 134/9720 [10:27<10:29:10,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.9828, 'learning_rate': 8.475298095115107e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                            | 135/9720 [10:30<10:17:22,  3.86s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1734, 'learning_rate': 8.540998080348558e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                            | 136/9720 [10:34<10:05:45,  3.79s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.8288, 'learning_rate': 8.606698065582008e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                             | 137/9720 [10:38<9:59:26,  3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.073, 'learning_rate': 8.672398050815458e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                            | 138/9720 [10:42<10:14:13,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.9704, 'learning_rate': 8.73809803604891e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                             | 139/9720 [10:45<9:59:18,  3.75s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.773, 'learning_rate': 8.80379802128236e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█                                                                             | 140/9720 [10:49<9:46:03,  3.67s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.7891, 'learning_rate': 8.86949800651581e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█▏                                                                            | 141/9720 [10:52<9:35:53,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.697, 'learning_rate': 8.935197991749261e-05, 'epoch': 0.04}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█▏                                                                            | 142/9720 [10:56<9:24:08,  3.53s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.7628, 'learning_rate': 9.000897976982711e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█▏                                                                            | 143/9720 [10:59<9:14:44,  3.48s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.7288, 'learning_rate': 9.066597962216162e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█▏                                                                            | 144/9720 [11:02<9:05:34,  3.42s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.7362, 'learning_rate': 9.132297947449613e-05, 'epoch': 0.04}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  1%|█▏                                                                            | 145/9720 [11:05<8:52:30,  3.34s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.6183, 'learning_rate': 9.197997932683063e-05, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                            | 146/9720 [11:08<8:36:32,  3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.4819, 'learning_rate': 9.263697917916512e-05, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                            | 147/9720 [11:11<8:21:32,  3.14s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.4482, 'learning_rate': 9.329397903149962e-05, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                            | 148/9720 [11:14<8:07:17,  3.05s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.0363, 'learning_rate': 9.395097888383412e-05, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                            | 149/9720 [11:17<7:53:33,  2.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 2.9486, 'learning_rate': 9.460797873616863e-05, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                            | 150/9720 [11:20<7:54:51,  2.98s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 6.0248, 'learning_rate': 9.526497858850313e-05, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                           | 151/9720 [11:28<11:49:40,  4.45s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                           | 152/9720 [11:35<13:48:57,  5.20s/it]
+  2%|█▏                                                                           | 152/9720 [11:35<13:48:57,  5.20s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.1906, 'learning_rate': 9.657897829317215e-05, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                           | 153/9720 [11:41<14:55:18,  5.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 5.0286, 'learning_rate': 9.723597814550665e-05, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                           | 154/9720 [11:48<15:27:51,  5.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.7783, 'learning_rate': 9.789297799784115e-05, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                           | 155/9720 [11:54<15:41:10,  5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5936, 'learning_rate': 9.854997785017566e-05, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                           | 156/9720 [12:00<15:44:34,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.7536, 'learning_rate': 9.920697770251016e-05, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▏                                                                           | 157/9720 [12:06<15:40:42,  5.90s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 158/9720 [12:11<15:26:04,  5.81s/it]
+  2%|█▎                                                                           | 158/9720 [12:11<15:26:04,  5.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.557, 'learning_rate': 0.00010052097740717918, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 159/9720 [12:17<15:10:11,  5.71s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.4231, 'learning_rate': 0.00010117797725951368, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 160/9720 [12:22<14:56:08,  5.62s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 161/9720 [12:27<14:42:25,  5.54s/it]
+  2%|█▎                                                                           | 161/9720 [12:27<14:42:25,  5.54s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4743, 'learning_rate': 0.0001024919769641827, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 162/9720 [12:33<14:27:09,  5.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4708, 'learning_rate': 0.0001031489768165172, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 163/9720 [12:38<14:30:37,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 164/9720 [12:43<14:13:54,  5.36s/it]
+  2%|█▎                                                                           | 164/9720 [12:43<14:13:54,  5.36s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5357, 'learning_rate': 0.00010446297652118621, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 165/9720 [12:48<13:56:19,  5.25s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 166/9720 [12:53<13:42:09,  5.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2262, 'learning_rate': 0.00010511997637352071, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.6843, 'learning_rate': 0.00010577697622585521, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 167/9720 [12:58<13:29:36,  5.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.3731, 'learning_rate': 0.00010643397607818973, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 168/9720 [13:03<13:14:24,  4.99s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 169/9720 [13:08<13:03:34,  4.92s/it]
+  2%|█▎                                                                           | 169/9720 [13:08<13:03:34,  4.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.2101, 'learning_rate': 0.00010774797578285873, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 170/9720 [13:12<12:54:07,  4.86s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.1394, 'learning_rate': 0.00010840497563519324, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 171/9720 [13:17<12:41:27,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 172/9720 [13:21<12:28:56,  4.71s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.2254, 'learning_rate': 0.00010906197548752774, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▎                                                                           | 173/9720 [13:26<12:17:59,  4.64s/it]
+  2%|█▎                                                                           | 173/9720 [13:26<12:17:59,  4.64s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.0966, 'learning_rate': 0.00011037597519219676, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 174/9720 [13:30<12:08:19,  4.58s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.3946, 'learning_rate': 0.00011103297504453126, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 175/9720 [13:35<12:15:55,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 176/9720 [13:39<12:02:58,  4.55s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1662, 'learning_rate': 0.00011168997489686576, 'epoch': 0.05}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 177/9720 [13:44<11:48:53,  4.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.3428, 'learning_rate': 0.00011234697474920027, 'epoch': 0.05}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 178/9720 [13:48<11:35:24,  4.37s/it]
+  2%|█▍                                                                           | 178/9720 [13:48<11:35:24,  4.37s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 179/9720 [13:52<11:22:11,  4.29s/it]
+  2%|█▍                                                                           | 179/9720 [13:52<11:22:11,  4.29s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 180/9720 [13:56<11:10:45,  4.22s/it]
+  2%|█▍                                                                           | 180/9720 [13:56<11:10:45,  4.22s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 181/9720 [14:00<11:01:01,  4.16s/it]
+  2%|█▍                                                                           | 181/9720 [14:00<11:01:01,  4.16s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 182/9720 [14:04<10:51:13,  4.10s/it]
+  2%|█▍                                                                           | 182/9720 [14:04<10:51:13,  4.10s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 183/9720 [14:08<10:41:57,  4.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.057, 'learning_rate': 0.00011628897386320728, 'epoch': 0.06}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 184/9720 [14:12<10:30:27,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.1179, 'learning_rate': 0.00011694597371554178, 'epoch': 0.06}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 185/9720 [14:16<10:23:27,  3.92s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1351, 'learning_rate': 0.00011760297356787629, 'epoch': 0.06}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.1048, 'learning_rate': 0.00011825997342021079, 'epoch': 0.06}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 186/9720 [14:19<10:16:12,  3.88s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 4.0181, 'learning_rate': 0.00011891697327254529, 'epoch': 0.06}
+  2%|█▍                                                                           | 187/9720 [14:23<10:06:48,  3.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.973, 'learning_rate': 0.0001195739731248798, 'epoch': 0.06}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 188/9720 [14:27<10:26:39,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.9068, 'learning_rate': 0.0001202309729772143, 'epoch': 0.06}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▍                                                                           | 189/9720 [14:31<10:17:14,  3.89s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▌                                                                            | 190/9720 [14:34<9:59:24,  3.77s/it]
+  2%|█▌                                                                            | 190/9720 [14:34<9:59:24,  3.77s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▌                                                                            | 191/9720 [14:38<9:44:09,  3.68s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.9339, 'learning_rate': 0.00012154497268188332, 'epoch': 0.06}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.6721, 'learning_rate': 0.00012220197253421782, 'epoch': 0.06}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▌                                                                            | 192/9720 [14:41<9:29:18,  3.59s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+{'loss': 3.7949, 'learning_rate': 0.00012285897238655232, 'epoch': 0.06}
+  2%|█▌                                                                            | 193/9720 [14:45<9:18:21,  3.52s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▌                                                                            | 194/9720 [14:48<9:06:20,  3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.3948, 'learning_rate': 0.00012351597223888682, 'epoch': 0.06}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▌                                                                            | 195/9720 [14:51<8:51:08,  3.35s/it]
+  2%|█▌                                                                            | 195/9720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+  2%|█▌                                                                            | 196/9720 [14:54<8:37:05,  3.26s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.228, 'learning_rate': 0.00012482997194355585, 'epoch': 0.06}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.2732, 'learning_rate': 0.00012548697179589035, 'epoch': 0.06}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.3839, 'learning_rate': 0.00012614397164822485, 'epoch': 0.06}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 2.9711, 'learning_rate': 0.00012680097150055935, 'epoch': 0.06}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 2.9525, 'learning_rate': 0.00012745797135289385, 'epoch': 0.06}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 6.2523, 'learning_rate': 0.00012811497120522835, 'epoch': 0.06}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [14:51<8:51:08,  3.35s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▌                                                                           | 202/9720 [15:21<13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▌                                                                           | 202/9720 [15:21<13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.485, 'learning_rate': 0.00012877197105756288, 'epoch': 0.06}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.9817, 'learning_rate': 0.00013008597076223188, 'epoch': 0.06}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5802, 'learning_rate': 0.00013074297061456638, 'epoch': 0.06}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:55:52,  5.27s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▋                                                                           | 206/9720 [15:46<15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5406, 'learning_rate': 0.0001327139701715699, 'epoch': 0.06}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.7411, 'learning_rate': 0.00013337097002390441, 'epoch': 0.06}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3877, 'learning_rate': 0.0001340279698762389, 'epoch': 0.06}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4414, 'learning_rate': 0.0001346849697285734, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:44:29,  5.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▋                                                                           | 212/9720 [16:19<14:27:35,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▋                                                                           | 212/9720 [16:19<14:27:35,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [16:19<14:27:35,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:27:35,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:27:35,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▋                                                                           | 213/9720 [16:25<14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.34, 'learning_rate': 0.00013731296913791142, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2321, 'learning_rate': 0.00013862696884258042, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5204, 'learning_rate': 0.00014256896795658745, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3385, 'learning_rate': 0.00014322596780892198, 'epoch': 0.07}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:32:21,  5.51s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▊                                                                           | 225/9720 [17:22<12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4033, 'learning_rate': 0.00014519696736592548, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2058, 'learning_rate': 0.00014585396721825998, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.313, 'learning_rate': 0.00014651096707059448, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2648, 'learning_rate': 0.00014716796692292898, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3322, 'learning_rate': 0.0001478249667752635, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3424, 'learning_rate': 0.0001491389664799325, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.9832, 'learning_rate': 0.000149795966332267, 'epoch': 0.07}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2187, 'learning_rate': 0.0001504529661846015, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:47,  4.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▊                                                                           | 236/9720 [18:06<10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.0884, 'learning_rate': 0.00015176696588927054, 'epoch': 0.07}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.7718, 'learning_rate': 0.00015439496529860854, 'epoch': 0.07}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1137, 'learning_rate': 0.00015505196515094304, 'epoch': 0.07}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.8398, 'learning_rate': 0.00015570896500327754, 'epoch': 0.07}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.7475, 'learning_rate': 0.00015636596485561204, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.3048, 'learning_rate': 0.00015702296470794654, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.4238, 'learning_rate': 0.00015833696441261557, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.1184, 'learning_rate': 0.00015899396426495007, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:08:14,  3.85s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██                                                                            | 250/9720 [18:52<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [18:52<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [18:52<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.7254, 'learning_rate': 0.00016096496382195358, 'epoch': 0.08}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.2385, 'learning_rate': 0.00016162196367428808, 'epoch': 0.08}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.9273, 'learning_rate': 0.0001622789635266226, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.7786, 'learning_rate': 0.0001635929632312916, 'epoch': 0.08}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4117, 'learning_rate': 0.0001642499630836261, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6439, 'learning_rate': 0.0001655639627882951, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4627, 'learning_rate': 0.0001662209626406296, 'epoch': 0.08}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4111, 'learning_rate': 0.00016687796249296414, 'epoch': 0.08}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4795, 'learning_rate': 0.00016819196219763314, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3954, 'learning_rate': 0.00016884896204996764, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.517, 'learning_rate': 0.00016950596190230214, 'epoch': 0.08}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4148, 'learning_rate': 0.00017016296175463664, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.707, 'learning_rate': 0.00017081996160697117, 'epoch': 0.08}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5414, 'learning_rate': 0.00017147696145930567, 'epoch': 0.08}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6632, 'learning_rate': 0.00017279096116397467, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3069, 'learning_rate': 0.00017344796101630917, 'epoch': 0.08}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:12,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▏                                                                          | 271/9720 [20:50<12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4037, 'learning_rate': 0.0001760759604256472, 'epoch': 0.08}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:38:42,  4.82s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▏                                                                          | 275/9720 [21:09<12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6712, 'learning_rate': 0.00017870395983498523, 'epoch': 0.09}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4491, 'learning_rate': 0.00017936095968731973, 'epoch': 0.09}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.0762, 'learning_rate': 0.00018001795953965423, 'epoch': 0.09}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2759, 'learning_rate': 0.00018067495939198873, 'epoch': 0.09}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.0968, 'learning_rate': 0.00018133195924432323, 'epoch': 0.09}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:12:08,  4.65s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▏                                                                          | 284/9720 [21:45<10:23:17,  3.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▏                                                                          | 284/9720 [21:45<10:23:17,  3.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [21:45<10:23:17,  3.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:23:17,  3.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:23:17,  3.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.9741, 'learning_rate': 0.00018330295880132676, 'epoch': 0.09}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:23:17,  3.96s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▎                                                                          | 286/9720 [21:53<10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.9358, 'learning_rate': 0.00018461695850599576, 'epoch': 0.09}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.4115, 'learning_rate': 0.00018855895762000276, 'epoch': 0.09}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:04:14,  3.84s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▎                                                                           | 294/9720 [22:21<9:00:33,  3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [22:21<9:00:33,  3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [22:21<9:00:33,  3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.7864, 'learning_rate': 0.00018987295732467177, 'epoch': 0.09}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:00:33,  3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:00:33,  3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.2292, 'learning_rate': 0.00019052995717700627, 'epoch': 0.09}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:00:33,  3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:00:33,  3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:00:33,  3.44s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▍                                                                           | 298/9720 [22:33<8:09:49,  3.12s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:09:49,  3.12s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:09:49,  3.12s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:09:49,  3.12s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:09:49,  3.12s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▍                                                                           | 300/9720 [22:39<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [22:39<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [22:39<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.8341, 'learning_rate': 0.0001938149564386788, 'epoch': 0.09}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.1676, 'learning_rate': 0.0001944719562910133, 'epoch': 0.09}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.7374, 'learning_rate': 0.00019644295584801683, 'epoch': 0.09}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.838, 'learning_rate': 0.00019775695555268583, 'epoch': 0.09}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5544, 'learning_rate': 0.00019841395540502033, 'epoch': 0.1}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6132, 'learning_rate': 0.00019907095525735483, 'epoch': 0.1}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2333, 'learning_rate': 0.00019972795510968933, 'epoch': 0.1}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4807, 'learning_rate': 0.00020038495496202386, 'epoch': 0.1}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4115, 'learning_rate': 0.00020104195481435836, 'epoch': 0.1}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.0106, 'learning_rate': 0.00020169895466669286, 'epoch': 0.1}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5739, 'learning_rate': 0.00020235595451902736, 'epoch': 0.1}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4143, 'learning_rate': 0.00020301295437136186, 'epoch': 0.1}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5588, 'learning_rate': 0.00020366995422369636, 'epoch': 0.1}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4126, 'learning_rate': 0.0002069549534853689, 'epoch': 0.1}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:03,  3.04s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▌                                                                          | 324/9720 [24:49<11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4859, 'learning_rate': 0.00020958295289470692, 'epoch': 0.1}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.273, 'learning_rate': 0.00021023995274704142, 'epoch': 0.1}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5106, 'learning_rate': 0.00021089695259937592, 'epoch': 0.1}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.048, 'learning_rate': 0.00021155395245171042, 'epoch': 0.1}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2437, 'learning_rate': 0.00021221095230404492, 'epoch': 0.1}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1273, 'learning_rate': 0.00021286795215637945, 'epoch': 0.1}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3951, 'learning_rate': 0.00021352495200871395, 'epoch': 0.1}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1273, 'learning_rate': 0.00021418195186104845, 'epoch': 0.1}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1578, 'learning_rate': 0.00021615295141805195, 'epoch': 0.1}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:54:14,  4.56s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▋                                                                           | 336/9720 [25:38<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1243, 'learning_rate': 0.00021812395097505548, 'epoch': 0.1}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:56:10,  3.81s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|██▋                                                                           | 341/9720 [25:56<9:25:11,  3.62s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:25:11,  3.62s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:25:11,  3.62s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.6799, 'learning_rate': 0.0002207519503843935, 'epoch': 0.11}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:25:11,  3.62s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:25:11,  3.62s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:25:11,  3.62s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:25:11,  3.62s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|██▊                                                                           | 344/9720 [26:06<8:56:01,  3.43s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [26:06<8:56:01,  3.43s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [26:06<8:56:01,  3.43s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:56:01,  3.43s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:56:01,  3.43s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|██▊                                                                           | 346/9720 [26:12<8:28:27,  3.25s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [26:12<8:28:27,  3.25s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [26:12<8:28:27,  3.25s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.4358, 'learning_rate': 0.00022403694964606602, 'epoch': 0.11}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [26:12<8:28:27,  3.25s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.371, 'learning_rate': 0.00022469394949840054, 'epoch': 0.11}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:28:27,  3.25s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:28:27,  3.25s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.3348, 'learning_rate': 0.00022535094935073504, 'epoch': 0.11}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:28:27,  3.25s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|██▊                                                                           | 350/9720 [26:24<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.1407, 'learning_rate': 0.00022797894876007302, 'epoch': 0.11}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.7007, 'learning_rate': 0.00022994994831707652, 'epoch': 0.11}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6645, 'learning_rate': 0.00023126394802174555, 'epoch': 0.11}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4648, 'learning_rate': 0.00023192094787408005, 'epoch': 0.11}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6883, 'learning_rate': 0.00023257794772641455, 'epoch': 0.11}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:51:43,  3.02s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|██▊                                                                          | 361/9720 [27:32<14:38:04,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [27:32<14:38:04,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [27:32<14:38:04,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [27:32<14:38:04,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5218, 'learning_rate': 0.00023389194743108355, 'epoch': 0.11}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [27:32<14:38:04,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [27:32<14:38:04,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:38:04,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:38:04,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:38:04,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:38:04,  5.63s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|██▉                                                                          | 364/9720 [27:48<14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [27:48<14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [27:48<14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4929, 'learning_rate': 0.00023783394654509058, 'epoch': 0.11}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3335, 'learning_rate': 0.0002398049461020941, 'epoch': 0.11}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:12:43,  5.47s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|██▉                                                                          | 372/9720 [28:27<12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4958, 'learning_rate': 0.00024243294551143214, 'epoch': 0.12}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.5047, 'learning_rate': 0.00024308994536376664, 'epoch': 0.12}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2801, 'learning_rate': 0.00024374694521610114, 'epoch': 0.12}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3793, 'learning_rate': 0.00024506094492077017, 'epoch': 0.12}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1101, 'learning_rate': 0.00024571794477310464, 'epoch': 0.12}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2544, 'learning_rate': 0.00024637494462543917, 'epoch': 0.12}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4169, 'learning_rate': 0.00024703194447777365, 'epoch': 0.12}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:24:20,  4.78s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|███                                                                          | 384/9720 [29:18<10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [29:18<10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [29:18<10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1355, 'learning_rate': 0.0002496599438871117, 'epoch': 0.12}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1599, 'learning_rate': 0.0002509739435917807, 'epoch': 0.12}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.6069, 'learning_rate': 0.00025294494314878423, 'epoch': 0.12}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.8343, 'learning_rate': 0.0002536019430011187, 'epoch': 0.12}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.8334, 'learning_rate': 0.00025425894285345323, 'epoch': 0.12}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:17:54,  3.97s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|███▏                                                                          | 394/9720 [29:54<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [29:54<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [29:54<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.5376, 'learning_rate': 0.00025557294255812224, 'epoch': 0.12}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [29:54<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:04:14,  3.50s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|███▏                                                                          | 400/9720 [30:12<7:57:55,  3.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:55,  3.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<7:57:55,  3.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:55,  3.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<7:57:55,  3.08s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|███▏                                                                         | 401/9720 [30:20<11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.8758, 'learning_rate': 0.00026148594122913277, 'epoch': 0.12}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 5.021, 'learning_rate': 0.0002621429410814673, 'epoch': 0.12}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6093, 'learning_rate': 0.00026279994093380177, 'epoch': 0.13}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.705, 'learning_rate': 0.0002634569407861363, 'epoch': 0.13}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.652, 'learning_rate': 0.00026411394063847077, 'epoch': 0.13}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.554, 'learning_rate': 0.0002647709404908053, 'epoch': 0.13}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4065, 'learning_rate': 0.0002654279403431398, 'epoch': 0.13}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4372, 'learning_rate': 0.0002660849401954743, 'epoch': 0.13}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3657, 'learning_rate': 0.00026674194004780883, 'epoch': 0.13}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4306, 'learning_rate': 0.00027068393916181583, 'epoch': 0.13}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4664, 'learning_rate': 0.0002713409390141503, 'epoch': 0.13}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3616, 'learning_rate': 0.00027331193857115383, 'epoch': 0.13}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3607, 'learning_rate': 0.0002739689384234883, 'epoch': 0.13}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2743, 'learning_rate': 0.00027528293812815736, 'epoch': 0.13}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3638, 'learning_rate': 0.00027659693783282636, 'epoch': 0.13}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.3029, 'learning_rate': 0.00027791093753749537, 'epoch': 0.13}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1737, 'learning_rate': 0.0002785679373898299, 'epoch': 0.13}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1625, 'learning_rate': 0.00027922493724216437, 'epoch': 0.13}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2028, 'learning_rate': 0.0002798819370944989, 'epoch': 0.13}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2445, 'learning_rate': 0.00028053893694683337, 'epoch': 0.13}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1621, 'learning_rate': 0.0002811959367991679, 'epoch': 0.13}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2919, 'learning_rate': 0.00028185293665150237, 'epoch': 0.13}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed11:49:04,  4.57s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|███▍                                                                         | 436/9720 [33:15<10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1263, 'learning_rate': 0.0002838239362085059, 'epoch': 0.14}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.0029, 'learning_rate': 0.0002844809360608404, 'epoch': 0.14}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed10:09:33,  3.94s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▌                                                                          | 442/9720 [33:38<9:27:37,  3.67s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:27:37,  3.67s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:27:37,  3.67s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:27:37,  3.67s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:27:37,  3.67s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.8678, 'learning_rate': 0.00028776593532251296, 'epoch': 0.14}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:27:37,  3.67s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:27:37,  3.67s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▌                                                                          | 445/9720 [33:48<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [33:48<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [33:48<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.6168, 'learning_rate': 0.00028907993502718196, 'epoch': 0.14}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [33:48<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.3913, 'learning_rate': 0.0002910509345841855, 'epoch': 0.14}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 6.6536, 'learning_rate': 0.0002923649342888545, 'epoch': 0.14}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:54:56,  3.46s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▌                                                                         | 453/9720 [34:25<14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.9266, 'learning_rate': 0.000295649933550527, 'epoch': 0.14}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed14:40:31,  5.70s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▋                                                                         | 458/9720 [34:55<15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [34:55<15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [34:55<15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [34:55<15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6102, 'learning_rate': 0.000298277932959865, 'epoch': 0.14}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.6471, 'learning_rate': 0.00029893493281219955, 'epoch': 0.14}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2794, 'learning_rate': 0.000299591932664534, 'epoch': 0.14}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.8004, 'learning_rate': 0.00030024893251686855, 'epoch': 0.14}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed15:14:41,  5.93s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▋                                                                         | 464/9720 [35:28<14:06:35,  5.49s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [35:28<14:06:35,  5.49s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [35:28<14:06:35,  5.49s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:06:35,  5.49s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:06:35,  5.49s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:06:35,  5.49s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed14:06:35,  5.49s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▋                                                                         | 466/9720 [35:38<13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [35:38<13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [35:38<13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4579, 'learning_rate': 0.0003035339317785411, 'epoch': 0.14}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4382, 'learning_rate': 0.0003068189310402136, 'epoch': 0.15}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed13:33:41,  5.28s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▊                                                                         | 475/9720 [36:21<12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [36:21<12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...20 [36:21<12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4076, 'learning_rate': 0.0003087899305972171, 'epoch': 0.15}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.279, 'learning_rate': 0.00030944693044955156, 'epoch': 0.15}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.4649, 'learning_rate': 0.0003101039303018861, 'epoch': 0.15}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.2617, 'learning_rate': 0.00031076093015422056, 'epoch': 0.15}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1989, 'learning_rate': 0.0003114179300065551, 'epoch': 0.15}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1309, 'learning_rate': 0.00031207492985888956, 'epoch': 0.15}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.254, 'learning_rate': 0.0003127319297112241, 'epoch': 0.15}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 4.1503, 'learning_rate': 0.0003140459294158931, 'epoch': 0.15}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed12:03:21,  4.69s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▉                                                                          | 492/9720 [37:27<9:15:27,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [37:27<9:15:27,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...720 [37:27<9:15:27,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:15:27,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:15:27,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.8771, 'learning_rate': 0.00032061592793923815, 'epoch': 0.15}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:15:27,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<9:15:27,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:15:27,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:15:27,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.3559, 'learning_rate': 0.00032192992764390715, 'epoch': 0.15}
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:15:27,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<9:15:27,  3.61s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▉                                                                          | 497/9720 [37:43<8:17:37,  3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▉                                                                          | 497/9720 [37:43<8:17:37,  3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Could not estimate the number of tokens of the input, floating-point operations will not be computed<8:17:37,  3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:17:37,  3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:17:37,  3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.0908, 'learning_rate': 0.0003239009272009107, 'epoch': 0.15}
+`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...e computed<8:17:37,  3.24s/it]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+{'loss': 3.1046, 'learning_rate': 0.0003245579270532452, 'epoch': 0.15}
+  0%|▏                                                                                 | 4/1845 [00:02<20:18,  1.51it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  0%|▎                                                                                 | 6/1845 [00:03<22:21,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  0%|▍                                                                                 | 9/1845 [00:06<22:25,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  1%|▍                                                                                | 11/1845 [00:07<22:39,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  1%|▌                                                                                | 14/1845 [00:10<24:08,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  1%|▋                                                                                | 17/1845 [00:12<23:28,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  1%|▊                                                                                | 19/1845 [00:14<24:36,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  1%|▉                                                                                | 22/1845 [00:16<22:15,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  1%|█                                                                                | 24/1845 [00:17<23:26,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  1%|█▏                                                                               | 27/1845 [00:20<25:06,  1.21it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▎                                                                               | 29/1845 [00:22<26:03,  1.16it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▎                                                                               | 31/1845 [00:23<26:10,  1.15it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▍                                                                               | 34/1845 [00:26<24:15,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▌                                                                               | 37/1845 [00:28<23:41,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▋                                                                               | 39/1845 [00:30<24:39,  1.22it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▊                                                                               | 42/1845 [00:32<21:39,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  2%|█▉                                                                               | 45/1845 [00:34<23:25,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██                                                                               | 47/1845 [00:36<22:04,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▏                                                                              | 50/1845 [00:38<23:37,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▎                                                                              | 53/1845 [00:40<22:20,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▍                                                                              | 55/1845 [00:42<22:45,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▌                                                                              | 58/1845 [00:44<20:47,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▋                                                                              | 61/1845 [00:46<22:27,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  3%|██▊                                                                              | 64/1845 [00:49<22:52,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|██▉                                                                              | 66/1845 [00:50<22:49,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|███                                                                              | 69/1845 [00:52<23:13,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|███                                                                              | 71/1845 [00:54<23:24,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|███▏                                                                             | 74/1845 [00:56<22:32,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|███▍                                                                             | 77/1845 [00:59<21:34,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|███▍                                                                             | 79/1845 [01:00<22:00,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  4%|███▌                                                                             | 82/1845 [01:03<24:05,  1.22it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▋                                                                             | 85/1845 [01:05<21:07,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▊                                                                             | 87/1845 [01:06<22:52,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|███▉                                                                             | 90/1845 [01:09<21:57,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|████                                                                             | 93/1845 [01:11<22:56,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|████▏                                                                            | 95/1845 [01:13<22:49,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|████▎                                                                            | 98/1845 [01:15<21:10,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  5%|████▍                                                                           | 101/1845 [01:17<20:51,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  6%|████▌                                                                           | 104/1845 [01:19<22:24,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  6%|████▌                                                                           | 106/1845 [01:20<20:29,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  6%|████▋                                                                           | 109/1845 [01:23<21:42,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  6%|████▊                                                                           | 112/1845 [01:25<23:09,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  6%|████▉                                                                           | 114/1845 [01:27<23:12,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  6%|█████                                                                           | 117/1845 [01:29<20:49,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  7%|█████▏                                                                          | 120/1845 [01:31<22:29,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  7%|█████▎                                                                          | 122/1845 [01:33<22:27,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  7%|█████▍                                                                          | 125/1845 [01:35<21:40,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  7%|█████▌                                                                          | 128/1845 [01:37<20:28,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  7%|█████▋                                                                          | 131/1845 [01:39<21:00,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  7%|█████▊                                                                          | 133/1845 [01:41<21:30,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  7%|█████▉                                                                          | 136/1845 [01:43<22:16,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  8%|██████                                                                          | 139/1845 [01:46<21:44,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  8%|██████                                                                          | 141/1845 [01:47<23:28,  1.21it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  8%|██████▏                                                                         | 143/1845 [01:49<22:18,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  8%|██████▎                                                                         | 146/1845 [01:51<21:01,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  8%|██████▍                                                                         | 148/1845 [01:52<19:42,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  8%|██████▌                                                                         | 151/1845 [01:55<20:13,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  8%|██████▋                                                                         | 155/1845 [01:58<20:00,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  9%|██████▊                                                                         | 158/1845 [02:00<19:08,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  9%|██████▉                                                                         | 161/1845 [02:02<19:52,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  9%|███████                                                                         | 163/1845 [02:03<22:13,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  9%|███████▏                                                                        | 166/1845 [02:06<20:13,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  9%|███████▎                                                                        | 168/1845 [02:07<21:24,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  9%|███████▍                                                                        | 171/1845 [02:09<20:48,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+  9%|███████▌                                                                        | 174/1845 [02:12<20:18,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 10%|███████▋                                                                        | 177/1845 [02:14<19:35,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 10%|███████▊                                                                        | 180/1845 [02:16<18:56,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 10%|███████▉                                                                        | 182/1845 [02:17<20:49,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 10%|████████                                                                        | 185/1845 [02:20<20:23,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 10%|████████                                                                        | 187/1845 [02:21<22:33,  1.23it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 10%|████████▏                                                                       | 189/1845 [02:23<20:52,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 10%|████████▎                                                                       | 192/1845 [02:25<19:54,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 11%|████████▍                                                                       | 195/1845 [02:27<19:26,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 11%|████████▌                                                                       | 198/1845 [02:30<21:14,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 11%|████████▋                                                                       | 200/1845 [02:31<20:26,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 11%|████████▊                                                                       | 203/1845 [02:33<21:36,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 11%|████████▉                                                                       | 205/1845 [02:35<21:53,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 11%|█████████                                                                       | 208/1845 [02:37<20:05,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 11%|█████████▏                                                                      | 211/1845 [02:40<21:44,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 12%|█████████▏                                                                      | 213/1845 [02:41<20:47,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 12%|█████████▎                                                                      | 216/1845 [02:44<22:29,  1.21it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 12%|█████████▍                                                                      | 218/1845 [02:45<21:10,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 12%|█████████▌                                                                      | 221/1845 [02:47<20:08,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 12%|█████████▋                                                                      | 224/1845 [02:50<21:18,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 12%|█████████▊                                                                      | 227/1845 [02:52<19:56,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 12%|█████████▉                                                                      | 229/1845 [02:54<21:16,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 13%|██████████                                                                      | 232/1845 [02:56<18:42,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 13%|██████████▏                                                                     | 235/1845 [02:58<19:45,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 13%|██████████▎                                                                     | 237/1845 [02:59<20:09,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 13%|██████████▍                                                                     | 240/1845 [03:02<19:19,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 13%|██████████▌                                                                     | 243/1845 [03:04<20:03,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 13%|██████████▋                                                                     | 246/1845 [03:06<19:34,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 13%|██████████▊                                                                     | 249/1845 [03:08<19:23,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 14%|██████████▉                                                                     | 251/1845 [03:10<18:14,  1.46it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 14%|███████████                                                                     | 254/1845 [03:12<20:28,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 14%|███████████                                                                     | 256/1845 [03:14<21:58,  1.21it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 14%|███████████▏                                                                    | 259/1845 [03:16<20:06,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 14%|███████████▎                                                                    | 262/1845 [03:18<21:06,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 14%|███████████▍                                                                    | 264/1845 [03:20<21:54,  1.20it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 14%|███████████▌                                                                    | 267/1845 [03:22<20:04,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 15%|███████████▋                                                                    | 269/1845 [03:24<20:38,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 15%|███████████▊                                                                    | 272/1845 [03:26<20:34,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 15%|███████████▉                                                                    | 275/1845 [03:29<20:08,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 15%|████████████                                                                    | 277/1845 [03:30<20:11,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 15%|█████��██████▏                                                                   | 280/1845 [03:32<20:12,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 15%|████████████▎                                                                   | 283/1845 [03:35<20:34,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 16%|████████████▍                                                                   | 286/1845 [03:37<19:46,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 16%|████████████▌                                                                   | 289/1845 [03:39<19:15,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 16%|████████████▌                                                                   | 291/1845 [03:41<20:13,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 16%|████████████▋                                                                   | 294/1845 [03:43<18:56,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 16%|████████████▊                                                                   | 296/1845 [03:44<18:18,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 16%|█████████████                                                                   | 300/1845 [03:47<18:53,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 16%|█████████████                                                                   | 302/1845 [03:49<19:36,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 16%|█████████████▏                                                                  | 304/1845 [03:50<19:33,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 17%|█████████████▎                                                                  | 307/1845 [03:53<19:33,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 17%|█████████████▍                                                                  | 311/1845 [03:56<18:36,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 17%|█████████████▌                                                                  | 313/1845 [03:57<19:57,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 17%|█████████████▋                                                                  | 316/1845 [03:59<18:56,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 17%|█████████████▊                                                                  | 319/1845 [04:01<17:36,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 17%|█████████████▉                                                                  | 321/1845 [04:03<18:50,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 18%|██████████████                                                                  | 324/1845 [04:05<18:52,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 18%|██████████████▏                                                                 | 326/1845 [04:07<18:08,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 18%|██████████████▎                                                                 | 329/1845 [04:09<17:29,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 18%|██████████████▍                                                                 | 332/1845 [04:11<19:45,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 18%|██████████████▍                                                                 | 334/1845 [04:13<19:47,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 18%|██████████████▌                                                                 | 337/1845 [04:15<18:48,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 18%|██████████████▋                                                                 | 340/1845 [04:17<20:06,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 19%|██████████████▊                                                                 | 343/1845 [04:19<17:39,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 19%|███████████████                                                                 | 346/1845 [04:21<18:15,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 19%|███████████████                                                                 | 348/1845 [04:23<17:29,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 19%|███████████████▏                                                                | 351/1845 [04:25<19:42,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 19%|███████████████▎                                                                | 354/1845 [04:28<19:17,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 19%|███████████████▍                                                                | 356/1845 [04:29<20:06,  1.23it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 19%|███████████████▌                                                                | 359/1845 [04:31<17:35,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 20%|███████████████▋                                                                | 362/1845 [04:34<18:31,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 20%|███████████████▊                                                                | 364/1845 [04:35<18:17,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 20%|███████████████▉                                                                | 367/1845 [04:37<18:47,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 20%|████████████████                                                                | 370/1845 [04:40<17:13,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 20%|████████████████▏                                                               | 373/1845 [04:42<18:01,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 20%|████████████████▎                                                               | 375/1845 [04:43<18:18,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 20%|████████████████▍                                                               | 378/1845 [04:45<17:58,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 21%|████████████████▍                                                               | 380/1845 [04:47<19:47,  1.23it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 21%|████████████████▌                                                               | 383/1845 [04:50<19:16,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 21%|████████████████▋                                                               | 386/1845 [04:52<18:14,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 21%|████████████████▊                                                               | 389/1845 [04:54<16:21,  1.48it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 21%|████████████████▉                                                               | 392/1845 [04:56<16:40,  1.45it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 21%|█████████████████▏                                                              | 395/1845 [04:58<18:10,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 22%|█████████████████▏                                                              | 397/1845 [05:00<19:11,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 22%|█████████████████▎                                                              | 400/1845 [05:02<18:30,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 22%|█████████████████▍                                                              | 402/1845 [05:04<18:24,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 22%|█████████████████▌                                                              | 405/1845 [05:06<19:42,  1.22it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 22%|█████████████████▋                                                              | 408/1845 [05:08<17:14,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 22%|█████████████████▊                                                              | 411/1845 [05:10<17:35,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 22%|█████████████████▉                                                              | 413/1845 [05:12<18:27,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 23%|██████████████████                                                              | 416/1845 [05:14<18:08,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 23%|██████████████████▏                                                             | 419/1845 [05:16<15:43,  1.51it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 23%|██████████████████▎                                                             | 421/1845 [05:18<18:43,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 23%|██████████████████▍                                                             | 424/1845 [05:20<16:50,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 23%|██████████████████▍                                                             | 426/1845 [05:22<16:56,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 23%|██████████████████▌                                                             | 429/1845 [05:24<17:30,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 23%|██████████████████▋                                                             | 432/1845 [05:26<16:31,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 24%|██████████████████▊                                                             | 435/1845 [05:28<17:17,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 24%|██████████████████▉                                                             | 437/1845 [05:30<18:24,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 24%|███████████████████                                                             | 440/1845 [05:32<15:45,  1.49it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 24%|███████████████████▏                                                            | 443/1845 [05:34<17:51,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 24%|███████████████████▎                                                            | 446/1845 [05:36<18:04,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 24%|███████████████████▍                                                            | 449/1845 [05:38<15:46,  1.48it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 24%|███████████████████▌                                                            | 451/1845 [05:40<19:40,  1.18it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 25%|███████████████████▋                                                            | 454/1845 [05:42<17:05,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 25%|███████████████████▊                                                            | 456/1845 [05:44<17:54,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 25%|███████████████████▉                                                            | 459/1845 [05:46<16:56,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 25%|████████████████████                                                            | 462/1845 [05:48<17:24,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 25%|████████████████████                                                            | 464/1845 [05:50<16:51,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 25%|████████████████████▏                                                           | 467/1845 [05:52<15:53,  1.45it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 25%|████████████████████▍                                                           | 470/1845 [05:54<17:55,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 26%|████████████████████▌                                                           | 473/1845 [05:57<17:20,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 26%|████████████████████▌                                                           | 475/1845 [05:58<17:05,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 26%|████████████████████▋                                                           | 478/1845 [06:00<17:44,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 26%|████████████████████▊                                                           | 480/1845 [06:02<17:49,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 26%|████████████████████▉                                                           | 484/1845 [06:04<14:40,  1.55it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 26%|█████████████████████                                                           | 486/1845 [06:06<14:53,  1.52it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 27%|█████████████████████▏                                                          | 490/1845 [06:09<14:48,  1.53it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 27%|█████████████████████▎                                                          | 492/1845 [06:10<15:19,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 27%|█████████████████████▍                                                          | 495/1845 [06:12<15:51,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 27%|█████████████████████▌                                                          | 498/1845 [06:14<14:50,  1.51it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 27%|█████████████████████▋                                                          | 501/1845 [06:16<16:28,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 27%|█████████████████████▊                                                          | 503/1845 [06:18<18:12,  1.23it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 27%|██████���██████████████▉                                                          | 506/1845 [06:21<17:20,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 28%|██████████████████████                                                          | 509/1845 [06:23<16:04,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 28%|██████████████████████▏                                                         | 512/1845 [06:25<17:21,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 28%|██████████████████████▎                                                         | 514/1845 [06:27<17:05,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 28%|██████████████████████▍                                                         | 517/1845 [06:29<17:10,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 28%|██████████████████████▌                                                         | 520/1845 [06:31<15:21,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 28%|██████████████████████▋                                                         | 523/1845 [06:33<15:52,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 28%|██████████████████████▊                                                         | 525/1845 [06:34<15:21,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 29%|██████████████████████▉                                                         | 528/1845 [06:37<16:29,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 29%|███████████████████████                                                         | 531/1845 [06:39<16:00,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 29%|███████████████████████▏                                                        | 534/1845 [06:41<17:07,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 29%|███████████████████████▏                                                        | 536/1845 [06:43<16:57,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 29%|███████████████████████▎                                                        | 539/1845 [06:45<17:34,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 29%|███████████████████████▍                                                        | 541/1845 [06:47<17:28,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 29%|███████████████████████▌                                                        | 544/1845 [06:49<15:53,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 30%|███████████████████████▋                                                        | 547/1845 [06:51<15:19,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 30%|███████████████████████▊                                                        | 549/1845 [06:53<16:52,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 30%|███████████████████████▉                                                        | 552/1845 [06:55<16:28,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 30%|████████████████████████                                                        | 555/1845 [06:57<16:33,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 30%|████████████████████████▏                                                       | 558/1845 [07:00<15:39,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 30%|████████████████████████▎                                                       | 560/1845 [07:01<15:51,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 31%|████████████████████████▍                                                       | 563/1845 [07:03<16:12,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 31%|████████████████████████▌                                                       | 566/1845 [07:05<14:22,  1.48it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 31%|████████████████████████▋                                                       | 569/1845 [07:08<16:31,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 31%|████████████████████████▊                                                       | 571/1845 [07:09<16:46,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 31%|████████████████████████▉                                                       | 574/1845 [07:12<15:56,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 31%|█████████████████████████                                                       | 577/1845 [07:14<16:07,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 31%|█████████████████████████                                                       | 579/1845 [07:15<17:00,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 32%|█████████████████████████▎                                                      | 583/1845 [07:18<13:37,  1.54it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 32%|█████████████████████████▍                                                      | 586/1845 [07:20<13:37,  1.54it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 32%|█████████████████████████▌                                                      | 589/1845 [07:22<14:12,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 32%|█████████████████████████▋                                                      | 591/1845 [07:24<15:28,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 32%|█████████████████████████▊                                                      | 594/1845 [07:26<14:31,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 32%|█████████████████████████▉                                                      | 597/1845 [07:28<15:33,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 32%|█████████████████████████▉                                                      | 599/1845 [07:30<16:22,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 33%|██████████████████████████                                                      | 602/1845 [07:32<16:44,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 33%|██████████████████████████▏                                                     | 604/1845 [07:33<15:42,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 33%|██████████████████████████▎                                                     | 607/1845 [07:36<16:08,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 33%|██████████████████████████▍                                                     | 610/1845 [07:38<15:51,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 33%|██████████████████████████▌                                                     | 613/1845 [07:40<15:43,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 33%|██████████████████████████▋                                                     | 615/1845 [07:42<16:05,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 33%|██████████████████████████▊                                                     | 618/1845 [07:44<16:09,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 34%|██████████████████████████▉                                                     | 621/1845 [07:46<14:59,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 34%|███████████████████████████                                                     | 623/1845 [07:48<15:13,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 34%|███████████████████████████                                                     | 625/1845 [07:49<16:21,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 34%|███████████████████████████▏                                                    | 628/1845 [07:52<16:19,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 34%|███████████████████████████▎                                                    | 631/1845 [07:54<14:49,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 34%|███████████████████████████▍                                                    | 633/1845 [07:56<16:08,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 34%|███████████████████████████▌                                                    | 636/1845 [07:58<15:43,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 35%|███████████████████████████▋                                                    | 638/1845 [08:00<15:38,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 35%|███████████████████████████▊                                                    | 641/1845 [08:02<15:10,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 35%|███████████████████████████▉                                                    | 644/1845 [08:04<14:42,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 35%|████████████████████████████                                                    | 646/1845 [08:06<16:14,  1.23it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 35%|████████████████████████████▏                                                   | 649/1845 [08:08<14:13,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 35%|████████████████████████████▎                                                   | 652/1845 [08:10<14:10,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 36%|████████████████████████████▍                                                   | 655/1845 [08:12<15:26,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 36%|████████████████████████████▍                                                   | 657/1845 [08:14<14:31,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 36%|████████████████████████████▌                                                   | 660/1845 [08:16<15:03,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 36%|████████████████████████████▋                                                   | 663/1845 [08:18<14:47,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 36%|████████████████████████████▉                                                   | 666/1845 [08:20<14:29,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 36%|████████████████████████████▉                                                   | 668/1845 [08:22<14:07,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 36%|█████████████████████████████                                                   | 671/1845 [08:24<13:56,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 37%|█████████████████████████████▏                                                  | 674/1845 [08:26<14:06,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 37%|█████████████████████████████▎                                                  | 677/1845 [08:28<13:34,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 37%|█████████████████████████████▍                                                  | 680/1845 [08:30<14:08,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 37%|█████████████████████████████▌                                                  | 682/1845 [08:32<14:33,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 37%|█████████████████████████████▋                                                  | 685/1845 [08:34<14:50,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 37%|█████████████████████████████▊                                                  | 688/1845 [08:36<13:37,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 37%|█████████████████████████████▉                                                  | 691/1845 [08:38<13:12,  1.46it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 38%|██████████████████████████████                                                  | 694/1845 [08:41<15:06,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 38%|██████████████████████████████▏                                                 | 697/1845 [08:43<15:27,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 38%|██████████████████████████████▎                                                 | 699/1845 [08:45<14:02,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 38%|██████████████████████████████▍                                                 | 702/1845 [08:47<13:51,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 38%|██████████████████████████████▌                                                 | 705/1845 [08:49<14:22,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 38%|██████████████████████████████▋                                                 | 707/1845 [08:51<14:49,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 38%|██████████████████████████████▊                                                 | 710/1845 [08:53<14:36,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 39%|██████████████████████████████▉                                                 | 713/1845 [08:55<14:38,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 39%|███████████████████████████████                                                 | 716/1845 [08:57<14:00,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 39%|███████████████████████████████▏                                                | 719/1845 [08:59<13:40,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 39%|███████████████████████████████▎                                                | 722/1845 [09:01<12:39,  1.48it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 39%|███████████████████████████████▍                                                | 725/1845 [09:04<13:25,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 39%|███████████████████████████████▌                                                | 727/1845 [09:05<14:17,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 40%|███████████████████████████████▌                                                | 729/1845 [09:07<15:02,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 40%|███████████████████████████████▋                                                | 732/1845 [09:09<14:07,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 40%|███████████████████████████████▊                                                | 735/1845 [09:11<13:49,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 40%|████████████████████████████████                                                | 738/1845 [09:13<13:00,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 40%|████████████████████████████████                                                | 740/1845 [09:15<14:31,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 40%|████████████████████████████████▏                                               | 743/1845 [09:18<15:00,  1.22it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 40%|████████████████████████████████▎                                               | 746/1845 [09:20<13:21,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 41%|████████████████████████████████▍                                               | 748/1845 [09:21<13:14,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 41%|████████████████████████████████▌                                               | 751/1845 [09:24<13:47,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 41%|████████████████████████████████▋                                               | 754/1845 [09:26<13:29,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 41%|████████████████████████████████▊                                               | 756/1845 [09:27<14:08,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 41%|████████████████████████████████▉                                               | 759/1845 [09:30<13:48,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 41%|████████████████████████████████▉                                               | 761/1845 [09:31<14:18,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 41%|█████████████████████████████████▏                                              | 764/1845 [09:34<14:41,  1.23it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 42%|█████████████████████████████████▏                                              | 766/1845 [09:36<15:13,  1.18it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 42%|█████████████████████████████████▎                                              | 769/1845 [09:38<13:05,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 42%|█████████████████████████████████▍                                              | 772/1845 [09:40<12:06,  1.48it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 42%|█████████████████████████████████▌                                              | 775/1845 [09:42<13:21,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 42%|█████████████████████████████████▋                                              | 778/1845 [09:44<12:39,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 42%|█████████████████████████████████▊                                              | 781/1845 [09:46<13:26,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 42%|█████████████████████████████████▉                                              | 784/1845 [09:48<12:20,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 43%|██████████████████████████████████                                              | 786/1845 [09:50<13:23,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 43%|██████████████████████████████████▏                                             | 789/1845 [09:52<14:19,  1.23it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 43%|██████████████████████████████████▎                                             | 791/1845 [09:54<14:15,  1.23it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 43%|██████████████████████████████████▍                                             | 794/1845 [09:56<13:34,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 43%|██████████████████████████████████▌                                             | 796/1845 [09:58<14:04,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 43%|██████████████████████████████████▋                                             | 799/1845 [10:00<12:59,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 43%|██████████████████████████████████▊                                             | 802/1845 [10:02<13:10,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 44%|██████████████████████████████████▊                                             | 804/1845 [10:04<13:24,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 44%|██████████████████████████████████▉                                             | 807/1845 [10:06<13:07,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 44%|███████████████████████████████████                                             | 809/1845 [10:08<13:12,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 44%|███████████████████████████████████▏                                            | 812/1845 [10:10<12:17,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 44%|███████████████████████████████████▎                                            | 815/1845 [10:12<12:57,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 44%|███████████████████████████████████▍                                            | 817/1845 [10:14<12:26,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 44%|███████████████████████████████████▌                                            | 820/1845 [10:16<12:46,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 45%|███████████████████████████████████▋                                            | 823/1845 [10:18<12:16,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 45%|███████████████████████████████████▊                                            | 826/1845 [10:20<12:22,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 45%|███████████████████████████████████▉                                            | 829/1845 [10:22<12:09,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 45%|████████████████████████████████████                                            | 832/1845 [10:24<11:33,  1.46it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 45%|████████████████████████████████████▏                                           | 834/1845 [10:26<12:48,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 45%|████████████████████████████████████▎                                           | 837/1845 [10:28<13:10,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 45%|████████████████████████████████████▍                                           | 839/1845 [10:30<13:09,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 46%|████████████████████████████████████▌                                           | 842/1845 [10:32<12:39,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 46%|████████████████████████████████████▋                                           | 845/1845 [10:35<13:03,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 46%|████████████████████████████████████▋                                           | 847/1845 [10:36<12:57,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 46%|████████████████████████████████████▊                                           | 850/1845 [10:38<11:52,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 46%|████████████████████████████████████▉                                           | 853/1845 [10:41<12:21,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 46%|█████████████████████████████████████                                           | 856/1845 [10:43<11:47,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 47%|█████████████████████████████████████▏                                          | 859/1845 [10:45<10:27,  1.57it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 47%|█████████████████████████████████████▍                                          | 862/1845 [10:47<11:39,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 47%|█████████████████████████████████████▌                                          | 865/1845 [10:49<11:30,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 47%|█████████████████████████████████████▋                                          | 868/1845 [10:51<11:39,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 47%|█████████████████████████████████████▋                                          | 870/1845 [10:53<12:25,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 47%|█████████████████████████████████████▊                                          | 873/1845 [10:55<11:51,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 47%|█████████████████████████████████████▉                                          | 876/1845 [10:57<12:30,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 48%|██████████████████████████████████████                                          | 878/1845 [10:59<12:33,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 48%|██████████████████████████████████████▏                                         | 881/1845 [11:01<12:44,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 48%|██████████████████████████████████████▎                                         | 884/1845 [11:03<12:25,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 48%|██████████████████████████████████████▍                                         | 887/1845 [11:05<11:17,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 48%|██████████████████████████████████████▌                                         | 889/1845 [11:07<11:11,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 48%|██████████████████████████████████████▋                                         | 892/1845 [11:09<11:17,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 49%|██████████████████████████████████████▊                                         | 895/1845 [11:11<11:31,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 49%|██████████████████████████████████████▉                                         | 898/1845 [11:13<11:40,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 49%|███████████████████████████████████████                                         | 900/1845 [11:15<11:55,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 49%|███████████████████████████████████████▏                                        | 904/1845 [11:17<10:37,  1.48it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 49%|███████████████████████████████████████▎                                        | 906/1845 [11:19<11:59,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 49%|███████████████████████████████████████▍                                        | 909/1845 [11:21<11:39,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 49%|███████████████████████████████████████▌                                        | 912/1845 [11:24<11:32,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 50%|███████████████████████████████████████▋                                        | 914/1845 [11:25<10:51,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 50%|███████████████████████████████████████▊                                        | 917/1845 [11:27<11:09,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 50%|███████████████████████████████████████▉                                        | 920/1845 [11:29<11:29,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 50%|███████████████████████████████████████▉                                        | 922/1845 [11:31<11:49,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 50%|████████████████████████████████████████                                        | 925/1845 [11:33<11:14,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 50%|████████████████████████████████████████▏                                       | 928/1845 [11:35<10:51,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 50%|████████████████████████████████████████▎                                       | 931/1845 [11:38<11:05,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 51%|████████████████████████████████████████▍                                       | 934/1845 [11:40<11:29,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 51%|████████████████████████████████████████▌                                       | 936/1845 [11:41<11:55,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 51%|████████████████████████████████████████▋                                       | 939/1845 [11:43<10:34,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 51%|████████████████████████████████████████▊                                       | 942/1845 [11:46<11:21,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 51%|████████████████████████████████████████▉                                       | 945/1845 [11:48<10:47,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 51%|█████████████████████████████████████████                                       | 948/1845 [11:50<10:56,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 51%|█████████████████████████████████████████▏                                      | 950/1845 [11:51<10:15,  1.45it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 52%|█████████████████████████████████████████▎                                      | 953/1845 [11:54<10:44,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 52%|█████████████████████████████████████████▍                                      | 955/1845 [11:55<11:55,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 52%|█████████████████████████████████████████▌                                      | 958/1845 [11:58<11:18,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 52%|█████████████████████████████████████████▋                                      | 960/1845 [11:59<11:40,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 52%|█████████████████████████████████████████▊                                      | 963/1845 [12:02<11:05,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 52%|█████████████████████████████████████████▉                                      | 966/1845 [12:04<10:58,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 53%|██████████████████████████████████████████                                      | 969/1845 [12:06<10:59,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 53%|██████████████████████████████████████████                                      | 971/1845 [12:08<11:17,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 53%|██████████████████████████████████████████▏                                     | 974/1845 [12:09<09:54,  1.46it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 53%|██████████████████████████████████████████▎                                     | 977/1845 [12:11<08:54,  1.62it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 53%|██████████████████████████████████████████▍                                     | 980/1845 [12:14<09:53,  1.46it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 53%|██████████████████████████████████████████▌                                     | 983/1845 [12:16<09:29,  1.51it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 53%|██████████████████████████████████████████▊                                     | 986/1845 [12:18<10:33,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 54%|██████████████████████████████████████████▉                                     | 989/1845 [12:20<10:26,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 54%|██████████████████████████████████████████▉                                     | 991/1845 [12:22<10:56,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 54%|███████████████████████████████████████████                                     | 994/1845 [12:24<10:30,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 54%|███████████████████████████████████████████▏                                    | 997/1845 [12:26<10:00,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 54%|██████████████████████████████████████████▊                                    | 1000/1845 [12:28<10:10,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 54%|██████████████████████████████████████████▉                                    | 1003/1845 [12:30<09:50,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 54%|███████████████████████████████████████████                                    | 1005/1845 [12:32<10:20,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 55%|███████████████████████████████████████████▏                                   | 1008/1845 [12:34<09:52,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 55%|███████████████████████████████████████████▎                                   | 1011/1845 [12:36<10:26,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 55%|███████████████████████████████████████████▍                                   | 1014/1845 [12:38<09:15,  1.50it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 55%|███████████████████████████████████████████▌                                   | 1017/1845 [12:40<09:28,  1.46it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 55%|███████████████████████████████████████████▋                                   | 1020/1845 [12:42<09:54,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 55%|███████████████████████████████████████████▊                                   | 1022/1845 [12:44<10:53,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 56%|███████████████████████████████████████████▉                                   | 1025/1845 [12:46<10:09,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 56%|████████████████████████████████████████████                                   | 1028/1845 [12:48<08:26,  1.61it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 56%|████████████████████████████████████████████▏                                  | 1031/1845 [12:50<09:06,  1.49it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 56%|████████████████████████████████████████████▎                                  | 1034/1845 [12:52<10:10,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 56%|████████████████████████████████████████████▍                                  | 1037/1845 [12:55<10:06,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 56%|████████████████████████████████████████████▌                                  | 1040/1845 [12:57<08:57,  1.50it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 57%|████████████████████████████████████████████▋                                  | 1043/1845 [12:59<09:16,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 57%|████████████████████████████████████████████▋                                  | 1045/1845 [13:00<09:12,  1.45it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 57%|████████████████████████████████████████████▊                                  | 1048/1845 [13:03<11:13,  1.18it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 57%|█████████████████████████████████████████████                                  | 1051/1845 [13:05<09:37,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 57%|█████████████████████████████████████████████                                  | 1053/1845 [13:06<09:22,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 57%|█████████████��███████████████████████████████▏                                 | 1056/1845 [13:08<09:07,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 57%|█████████████████████████████████████████████▎                                 | 1059/1845 [13:11<09:38,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 58%|█████████████████████████████████████████████▍                                 | 1062/1845 [13:13<09:26,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 58%|█████████████████████████████████████████████▌                                 | 1065/1845 [13:15<09:00,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 58%|█████████████████████████████████████████████▋                                 | 1068/1845 [13:17<08:38,  1.50it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 58%|█████████████████████████████████████████████▊                                 | 1071/1845 [13:19<08:31,  1.51it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 58%|█████████████████████████████████████████████▉                                 | 1073/1845 [13:20<09:14,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 58%|██████████████████████████████████████████████                                 | 1076/1845 [13:23<09:21,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 58%|██████████████████████████████████████████████▏                                | 1079/1845 [13:25<08:35,  1.49it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 59%|██████████████████████████████████████████████▎                                | 1082/1845 [13:27<09:41,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 59%|██████████████████████████████████████████████▍                                | 1085/1845 [13:29<09:16,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 59%|██████████████████████████████████████████████▌                                | 1087/1845 [13:31<10:03,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 59%|██████████████████████████████████████████████▋                                | 1090/1845 [13:33<10:04,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 59%|██████████████████████████████████████████████▊                                | 1092/1845 [13:35<09:22,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 59%|██████████████████████████████████████████████▉                                | 1095/1845 [13:37<09:00,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 60%|███████████████████████████████████████████████                                | 1098/1845 [13:39<09:27,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 60%|███████████████████████████████████████████████▏                               | 1101/1845 [13:41<09:19,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 60%|███████████████████████████████████████████████▏                               | 1103/1845 [13:43<09:35,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 60%|███████████████████████████████████████████████▎                               | 1106/1845 [13:45<08:59,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 60%|███████████████████████████████████████████████▍                               | 1109/1845 [13:47<08:19,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 60%|███████████████████████████████████████████████▌                               | 1112/1845 [13:49<08:28,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 60%|███████████████████████████████████████████████▋                               | 1115/1845 [13:51<08:18,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 61%|███████████████████████████████████████████████▊                               | 1118/1845 [13:54<09:24,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 61%|███████████████████████████████████████████████▉                               | 1120/1845 [13:55<09:16,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 61%|████████████████████████████████████████████████                               | 1123/1845 [13:57<08:41,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 61%|████████████████████████████████████████████████▏                              | 1126/1845 [14:00<08:31,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 61%|████████████████████████████████████████████████▎                              | 1129/1845 [14:02<08:52,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 61%|████████████████████████████████████████████████▍                              | 1132/1845 [14:04<07:58,  1.49it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 62%|████████████████████████████████████████████████▌                              | 1135/1845 [14:06<08:11,  1.45it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 62%|████████████████████████████████████████████████▋                              | 1138/1845 [14:08<07:48,  1.51it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 62%|████████████████████████████████████████████████▊                              | 1140/1845 [14:09<07:33,  1.55it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 62%|████████████████████████████████████████████████▉                              | 1144/1845 [14:12<07:36,  1.54it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 62%|█████████████████████████████████████████████████                              | 1146/1845 [14:13<08:19,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 62%|█████████████████████████████████████████████████��                             | 1149/1845 [14:15<08:37,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 62%|█████████████████████████████████████████████████▎                             | 1152/1845 [14:18<08:42,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 63%|█████████████████████████████████████████████████▍                             | 1154/1845 [14:19<08:25,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 63%|█████████████████████████████████████████████████▌                             | 1157/1845 [14:21<08:21,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 63%|█████████████████████████████████████████████████▋                             | 1160/1845 [14:24<09:07,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 63%|█████████████████████████████████████████████████▊                             | 1163/1845 [14:26<08:32,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 63%|█████████████████████████████████████████████████▉                             | 1165/1845 [14:27<08:21,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 63%|██████████████████████████████████████████████████                             | 1168/1845 [14:30<09:07,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 63%|██████████████████████████████████████████████████                             | 1170/1845 [14:32<08:50,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 64%|██████████████████████████████████████████████████▏                            | 1173/1845 [14:34<08:59,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 64%|██████████████████████████████████████████████████▎                            | 1175/1845 [14:35<08:28,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 64%|██████████████████████████████████████████████████▍                            | 1178/1845 [14:38<08:04,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 64%|██████████████████████████████████████████████████▌                            | 1181/1845 [14:40<07:48,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 64%|██████████████████████████████████████████████████▋                            | 1184/1845 [14:42<08:22,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 64%|██████████████████████████████████████████████████▊                            | 1186/1845 [14:44<08:25,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 64%|██████████████████████████████████████████████████▉                            | 1189/1845 [14:46<07:40,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 65%|███████████████████████████████████████████████████                            | 1192/1845 [14:48<08:24,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 65%|███████████████████████████████████████████████████▏                           | 1194/1845 [14:50<08:17,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 65%|███████████████████████████████████████████████████▎                           | 1197/1845 [14:52<07:43,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 65%|███████████████████████████████████████████████████▍                           | 1200/1845 [14:54<07:31,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 65%|███████████████████████████████████████████████████▌                           | 1203/1845 [14:56<07:50,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 65%|███████████████████████████████████████████████████▋                           | 1206/1845 [14:58<07:45,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 66%|███████████████████████████████████████████████████▊                           | 1209/1845 [15:00<07:44,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 66%|███████████████████████████████████████████████████▊                           | 1211/1845 [15:02<08:04,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 66%|███████████████████████████████████████████████████▉                           | 1214/1845 [15:04<07:44,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 66%|████████████████████████████████████████████████████                           | 1217/1845 [15:06<07:05,  1.48it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 66%|████████████████████████████████████████████████████▏                          | 1220/1845 [15:08<07:50,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 66%|████████████████████████████████████████████████████▎                          | 1223/1845 [15:11<07:26,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 66%|████████████████████████████████████████████████████▍                          | 1226/1845 [15:13<07:21,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 67%|████████████████████████████████████████████████████▌                          | 1228/1845 [15:14<07:33,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 67%|████████████████████████████████████████████████████▋                          | 1231/1845 [15:16<07:05,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 67%|████████████████████████████████████████████████████▊                          | 1234/1845 [15:18<06:49,  1.49it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 67%|████████████████████████████████████████████████████▉                          | 1236/1845 [15:20<07:25,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 67%|█████████████████████████████████████████████████████                          | 1240/1845 [15:23<07:30,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 67%|█████████████████████████████████████████████████████▏                         | 1243/1845 [15:25<06:58,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 68%|█████████████████████████████████████████████████████▎                         | 1246/1845 [15:27<06:52,  1.45it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 68%|█████████████████████████████████████████████████████▍                         | 1248/1845 [15:29<07:06,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 68%|█████████████████████████████████████████████████████▌                         | 1251/1845 [15:31<06:59,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 68%|█████████████████████████████████████████████████████▋                         | 1254/1845 [15:33<07:47,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 68%|█████████████████████████████████████████████████████▊                         | 1256/1845 [15:35<07:33,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 68%|█████████████████████████████████████████████████████▊                         | 1258/1845 [15:36<07:23,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 68%|█████████████████████████████████████████████████████▉                         | 1261/1845 [15:38<07:26,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 68%|██████████████████████████████████████████████████████                         | 1263/1845 [15:40<07:40,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 69%|██████████████████████████████████████████████████████▏                        | 1266/1845 [15:43<07:59,  1.21it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 69%|██████████████████████████████████████████████████████▎                        | 1269/1845 [15:45<07:16,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 69%|██████████████████████████████████████████████████████▍                        | 1271/1845 [15:46<07:24,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 69%|██████████████████████████████████████████████████████▌                        | 1274/1845 [15:48<06:45,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 69%|██████████████████████████████████████████████████████▋                        | 1277/1845 [15:50<06:35,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 69%|██████████████████████████████████████████████████████▊                        | 1280/1845 [15:53<06:50,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 70%|██████████████████████████████████████████████████████▉                        | 1283/1845 [15:55<07:05,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 70%|███████████████████████████████████████████████████████                        | 1285/1845 [15:57<07:24,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 70%|███████████████████████████████████████████████████████▏                       | 1288/1845 [15:58<06:15,  1.48it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 70%|███████████████████████████████████████████████████████▎                       | 1291/1845 [16:01<06:42,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 70%|███████████████████████████████████████████████████████▍                       | 1294/1845 [16:03<06:34,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 70%|███████████████████████████████████████████████████████▍                       | 1296/1845 [16:05<07:07,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 70%|███████████████████████████████████████████████████████▌                       | 1299/1845 [16:07<06:51,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 71%|███████████████████████████████████████████████████████▋                       | 1301/1845 [16:08<06:58,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 71%|███████████████████████████████████████████████████████▊                       | 1304/1845 [16:11<06:37,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 71%|███████████████████████████████████████████████████████▉                       | 1307/1845 [16:13<06:43,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 71%|████████████████████████████████████████████████████████                       | 1309/1845 [16:15<06:59,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 71%|████████████████████████████████████████████████████████▏                      | 1312/1845 [16:17<07:03,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 71%|████████████████████████████████████████████████████████▎                      | 1314/1845 [16:19<07:07,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 71%|████████████████████████████████████████████████████████▍                      | 1317/1845 [16:21<07:00,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 72%|████████████████████████████████████████████████████████▌                      | 1320/1845 [16:23<06:18,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 72%|████████████████████████████████████████████████████████▋                      | 1323/1845 [16:25<06:40,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 72%|████████████████████████████████████████████████████████▋                      | 1325/1845 [16:27<06:26,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 72%|████████████████████████████████████████████████████████▊                      | 1328/1845 [16:29<06:53,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 72%|████████████████████████████████████████████████████████▉                      | 1331/1845 [16:32<06:30,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 72%|█████████████████████████████████████████████████████████                      | 1334/1845 [16:34<06:18,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 72%|█████████████████████████████████████████████████████████▏                     | 1336/1845 [16:35<06:48,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 73%|█████████████████████████████████████████████████████████▎                     | 1339/1845 [16:37<05:58,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 73%|█████████████████████████████████████████████████████████▍                     | 1342/1845 [16:39<06:03,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 73%|█████████████████████████████████████████████████████████▌                     | 1344/1845 [16:41<06:27,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 73%|█████████████████████████████████████████████████████████▋                     | 1348/1845 [16:44<05:36,  1.48it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 73%|█████████████████████████████████████████████████████████▊                     | 1351/1845 [16:46<05:37,  1.46it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 73%|█████████████████████████████████████████████████████████▉                     | 1354/1845 [16:48<05:33,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 74%|██████████████████████████████████████████████████████████                     | 1357/1845 [16:50<05:11,  1.57it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 74%|██████████████████████████████████████████████████████████▏                    | 1360/1845 [16:52<05:19,  1.52it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 74%|██████████████████████████████████████████████████████████▎                    | 1363/1845 [16:54<05:27,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 74%|██████████████████████████████████████████████████████████▌                    | 1367/1845 [16:57<05:40,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 74%|████████████████████████████████████████���█████████████████▋                    | 1370/1845 [16:59<05:14,  1.51it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 74%|██████████████████████████████████████████████████████████▊                    | 1373/1845 [17:01<05:20,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 75%|██████████████████████████████████████████████████████████▉                    | 1375/1845 [17:02<06:03,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 75%|███████████████████████████████████████████████████████████                    | 1378/1845 [17:04<05:32,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 75%|███████████████████████████████████████████████████████████▏                   | 1381/1845 [17:07<05:45,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 75%|███████████████████████████████████████████████████████████▎                   | 1384/1845 [17:09<05:39,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 75%|███████████████████████████████████████████████████████████▍                   | 1387/1845 [17:11<05:00,  1.52it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 75%|███████████████████████████████████████████████████████████▌                   | 1390/1845 [17:13<05:17,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 76%|███████████████████████████████████████████████████████████▋                   | 1393/1845 [17:15<05:08,  1.46it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 76%|███████████████████████████████████████████████████████████▊                   | 1396/1845 [17:17<05:01,  1.49it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 76%|███████████████████████████████████████████████████████████▊                   | 1398/1845 [17:18<05:06,  1.46it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 76%|███████████████████████████████████████████████████████████▉                   | 1401/1845 [17:21<05:36,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 76%|████████████████████████████████████████████████████████████                   | 1404/1845 [17:23<05:21,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 76%|████████████████████████████████████████████████████████████▏                  | 1407/1845 [17:25<05:19,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 76%|████████████████████████████████████████████████████████████▎                  | 1409/1845 [17:27<05:26,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 77%|████████████████████████████████████████████████████████████▍                  | 1412/1845 [17:29<05:16,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 77%|████████████████████████████████████████████████████████████▌                  | 1415/1845 [17:31<05:02,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 77%|████████████████████████████████████████████████████████████▋                  | 1418/1845 [17:33<04:24,  1.62it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 77%|████████████████████████████████████████████████████████████▊                  | 1420/1845 [17:34<04:58,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 77%|████████████████████████████████████████████████████████████▉                  | 1423/1845 [17:36<05:12,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 77%|█████████████████████████████████████████████████████████████                  | 1426/1845 [17:39<05:07,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 77%|█████████████████████████████████████████████████████████████▏                 | 1428/1845 [17:40<05:00,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 78%|█████████████████████████████████████████████████████████████▎                 | 1431/1845 [17:42<05:10,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 78%|█████████████████████████████████████████████████████████████▎                 | 1433/1845 [17:44<05:39,  1.21it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 78%|█████████████████████████████████████████████████████████████▍                 | 1436/1845 [17:46<05:03,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 78%|█████████████████████████████████████████████████████████████▌                 | 1439/1845 [17:48<04:46,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 78%|█████████████████████████████████████████████████████████████▋                 | 1442/1845 [17:51<04:49,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 78%|█████████████████████████████████████████████████████████████▊                 | 1445/1845 [17:53<04:41,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 78%|██████████████████████████████████████████████████████████████                 | 1448/1845 [17:55<04:42,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 79%|██████████████████████████████████████████████████████████████                 | 1450/1845 [17:56<04:51,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 79%|██████████████████████████████████████████████████████████████▏                | 1453/1845 [17:59<04:37,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 79%|██████���███████████████████████████████████████████████████████▎                | 1456/1845 [18:01<04:37,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 79%|██████████████████████████████████████████████████████████████▍                | 1459/1845 [18:03<05:12,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 79%|██████████████████████████████████████████████████████████████▌                | 1461/1845 [18:05<05:04,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 79%|██████████████████████████████████████████████████████████████▋                | 1465/1845 [18:08<04:45,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 80%|██████████████████████████████████████████████████████████████▊                | 1467/1845 [18:09<04:43,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 80%|██████████████████████████████████████████████████████████████▉                | 1470/1845 [18:11<04:38,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 80%|███████████████████████████████████████████████████████████████                | 1473/1845 [18:14<04:37,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 80%|███████████████████████████████████████████████████████████████▏               | 1475/1845 [18:15<04:21,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 80%|███████████████████████████████████████████████████████████████▎               | 1478/1845 [18:17<04:23,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 80%|███████████████████████████████████████████████████████████████▍               | 1481/1845 [18:20<04:31,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 80%|███████████████████████████████████████████████████████████████▌               | 1484/1845 [18:22<04:19,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 81%|███████████████████████████████████████████████████████████████▋               | 1487/1845 [18:24<04:15,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 81%|███████████████████████████████████████████████████████████████▊               | 1489/1845 [18:25<04:31,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 81%|███████████████████████████████████████████████████████████████▉               | 1492/1845 [18:28<04:25,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 81%|████████████████████████████████████████████████████████████████               | 1495/1845 [18:30<04:00,  1.45it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 81%|█████████████████��██████████████████████████████████████████████               | 1497/1845 [18:31<03:59,  1.45it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 81%|████████████████████████████████████████████████████████████████▏              | 1500/1845 [18:33<04:14,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 81%|████████████████████████████████████████████████████████████████▎              | 1502/1845 [18:35<04:32,  1.26it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 82%|████████████████████████████████████████████████████████████████▍              | 1505/1845 [18:37<04:25,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 82%|████████████████████████████████████████████████████████████████▌              | 1508/1845 [18:40<04:12,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 82%|████████████████████████████████████████████████████████████████▋              | 1510/1845 [18:41<04:31,  1.23it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 82%|████████████████████████████████████████████████████████████████▊              | 1513/1845 [18:43<04:14,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 82%|████████████████████████████████████████████████████████████████▉              | 1516/1845 [18:46<04:05,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 82%|█████████████████████████████████████████████████████████████████              | 1520/1845 [18:48<03:34,  1.51it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 83%|█████████████████████████████████████████████████████████████████▏             | 1523/1845 [18:50<03:46,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 83%|█████████████████████████████████████████████████████████████████▎             | 1525/1845 [18:52<03:42,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 83%|█████████████████████████████████████████████████████████████████▍             | 1528/1845 [18:54<03:40,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 83%|█████████████████████████████████████████████████████████████████▌             | 1531/1845 [18:56<03:28,  1.50it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 83%|█████████████████████████████████████████████████████████████████▋             | 1534/1845 [18:58<03:48,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 83%|█████████████████████████████████████████████████████████████████▊             | 1537/1845 [19:00<03:48,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 83%|█████████████████████████████████████████████████████████████████▉             | 1540/1845 [19:02<03:43,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 84%|██████████████████████████████████████████████████████████████████             | 1543/1845 [19:05<03:31,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 84%|██████████████████████████████████████████████████████████████████▏            | 1545/1845 [19:06<03:31,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 84%|██████████████████████████████████████████████████████████████████▎            | 1548/1845 [19:08<03:23,  1.46it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 84%|██████████████████████████████████████████████████████████████████▍            | 1551/1845 [19:10<03:32,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 84%|██████████████████████████████████████████████████████████████████▍            | 1553/1845 [19:12<03:45,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 84%|██████████████████████████████████████████████████████████████████▋            | 1557/1845 [19:15<03:27,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 84%|██████████████████████████████████████████████████████████████████▊            | 1559/1845 [19:16<03:28,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 85%|██████████████████████████████████████████████████████████████████▉            | 1562/1845 [19:18<03:16,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 85%|███████████████████████████████████████████████████████████████████            | 1565/1845 [19:20<03:23,  1.38it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 85%|███████████████████████████████████████████████████████████████████▏           | 1568/1845 [19:23<03:31,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 85%|███████████████████████████████████████████████████████████████████▏           | 1570/1845 [19:24<03:32,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 85%|███████████████████████████████████████████████████████████████████▎           | 1573/1845 [19:27<03:33,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 85%|███████████████████████████████████████████████████████████████████▍           | 1575/1845 [19:28<03:29,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 86%|███████████████████████████████████████████████████████████████████▌           | 1578/1845 [19:31<03:34,  1.25it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 86%|███████████████████████████████████████████████████████████████████▋           | 1581/1845 [19:33<03:04,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 86%|███████████████████████████████████████████████████████████████████▊           | 1584/1845 [19:35<03:12,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 86%|███████████████████████████████████████████████████████████████████▉           | 1586/1845 [19:37<03:16,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 86%|████████████████████████████████████████████████████████████████████           | 1589/1845 [19:39<03:29,  1.22it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 86%|████████████████████████████████████████████████████████████████████▏          | 1592/1845 [19:41<03:08,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 86%|████████████████████████████████████████████████████████████████████▎          | 1595/1845 [19:43<02:53,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 87%|████████████████████████████████████████████████████████████████████▍          | 1598/1845 [19:45<02:47,  1.48it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 87%|████████████████████████████████████████████████████████████████████▌          | 1601/1845 [19:47<02:43,  1.49it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 87%|████████████████████████████████████████████████████████████████████▋          | 1603/1845 [19:49<02:40,  1.51it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 87%|████████████████████████████████████████████████████████████████████▊          | 1606/1845 [19:51<02:54,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 87%|████████████████████████████████████████████████████████████████████▉          | 1609/1845 [19:53<02:44,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 87%|█████████████████████████████████████████████████████████████████████          | 1612/1845 [19:55<02:57,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 87%|█████████████████████████████████████████████████████████████████████          | 1614/1845 [19:57<02:53,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 88%|█████████████████████████████████████████████████████████████████████▏         | 1617/1845 [19:59<02:38,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 88%|█████████████████████████████████████████████████████████████████████▎         | 1620/1845 [20:01<02:47,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 88%|█████████████████████████████████████████████████████████████████████▍         | 1622/1845 [20:03<03:05,  1.20it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 88%|█████████████████████████████████████████████████████████████████████▌         | 1625/1845 [20:05<02:45,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 88%|█████████████████████████████████████████████████████████████████████▋         | 1628/1845 [20:07<02:41,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 88%|█████████████████████████████████████████████████████████████████████▊         | 1630/1845 [20:09<02:53,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 89%|█████████████████████████████████████████████████████████████████████▉         | 1633/1845 [20:11<02:32,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 89%|██████████████████████████████████████████████████████████████████████         | 1636/1845 [20:13<02:22,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 89%|██████████████████████████████████████████████████████████████████████▏        | 1639/1845 [20:15<02:24,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 89%|██████████████████████████████████████████████████████████████████████▎        | 1642/1845 [20:18<02:38,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 89%|██████████████████████████████████████████████████████████████████████▍        | 1644/1845 [20:19<02:37,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 89%|██████████████████████████████████████████████████████████████████████▌        | 1647/1845 [20:22<02:32,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 89%|██████████████████████████████████████████████████████████████████████▌        | 1649/1845 [20:23<02:30,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 90%|██████████████████████████████████████████████████████████████████████▊        | 1653/1845 [20:26<02:11,  1.46it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 90%|██████████████████████████████████████████████████████████████████████▊        | 1655/1845 [20:27<02:24,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 90%|███████████████████████████████████████████████████████████████████████        | 1659/1845 [20:30<02:00,  1.54it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 90%|██���████████████████████████████████████████████████████████████████████        | 1661/1845 [20:31<02:03,  1.49it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 90%|███████████████████████████████████████████████████████████████████████▏       | 1664/1845 [20:33<02:02,  1.48it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 90%|███████████████████████████████████████████████████████████████████████▍       | 1667/1845 [20:36<02:12,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 91%|███████████████████████████████████████████████████████████████████████▌       | 1670/1845 [20:38<02:12,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 91%|███████████████████████████████████████████████████████████████████████▌       | 1672/1845 [20:40<02:15,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 91%|███████████████████████████████████████████████████████████████████████▋       | 1675/1845 [20:42<02:07,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 91%|███████████████████████████████████████████████████████████████████████▊       | 1678/1845 [20:44<02:08,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 91%|███████████████████████████████████████████████████████████████████████▉       | 1681/1845 [20:46<02:00,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 91%|████████████████████████████████████████████████████████████████████████       | 1684/1845 [20:48<01:53,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 91%|████████████████████████████████████████████████████████████████████████▏      | 1686/1845 [20:50<01:59,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 92%|████████████████████████████████████████████████████████████████████████▎      | 1690/1845 [20:52<01:34,  1.63it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 92%|████████████████████████████████████████████████████████████████████████▍      | 1693/1845 [20:54<01:41,  1.49it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 92%|████████████████████████████████████████████████████████████████████████▌      | 1695/1845 [20:56<01:49,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 92%|████████████████████████████████████████████████████████████████████████▋      | 1698/1845 [20:58<01:48,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 92%|████████████████████████████████████████████████████████████████████████▊      | 1701/1845 [21:00<01:42,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 92%|████████████████████████████████████████████████████████████████████████▉      | 1703/1845 [21:02<01:48,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 92%|█████████████████████████████████████████████████████████████████████████      | 1706/1845 [21:04<01:47,  1.30it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 93%|█████████████████████████████████████████████████████████████████████████▏     | 1709/1845 [21:06<01:33,  1.45it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 93%|█████████████████████████████████████████████████████████████████████████▎     | 1712/1845 [21:09<01:34,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 93%|█████████████████████████████████████████████████████████████████████████▍     | 1715/1845 [21:11<01:33,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 93%|█████████████████████████████████████████████████████████████████████████▌     | 1717/1845 [21:12<01:33,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 93%|█████████████████████████████████████████████████████████████████████████▋     | 1720/1845 [21:14<01:32,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 93%|█████████████████████████████████████████████████████████████████████████▊     | 1723/1845 [21:16<01:27,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 93%|█████████████████████████████████████████████████████████████████████████▊     | 1725/1845 [21:18<01:28,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 94%|█████████████████████████████████████████████████████████████████████████▉     | 1728/1845 [21:20<01:26,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 94%|██████████████████████████████████████████████████████████████████████████     | 1731/1845 [21:22<01:19,  1.43it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 94%|██████████████████████████████████████████████████████████████████████████▏    | 1733/1845 [21:24<01:19,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 94%|██████████████████████████████████████████████████████████████████████████▎    | 1736/1845 [21:26<01:18,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 94%|████████████████████████████████████████████████████████████████████████���█▍    | 1739/1845 [21:28<01:17,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 94%|██████████████████████████████████████████████████████████████████████████▋    | 1743/1845 [21:31<01:09,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 95%|██████████████████████████████████████████████████████████████████████████▊    | 1746/1845 [21:33<01:08,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 95%|██████████████████████████████████████████████████████████████████████████▉    | 1749/1845 [21:35<01:00,  1.59it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 95%|██████████████████████████████████████████████████████████████████████████▉    | 1751/1845 [21:36<01:01,  1.52it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 95%|███████████████████████████████████████████████████████████████████████████    | 1754/1845 [21:39<01:13,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 95%|███████████████████████████████████████████████████████████████████████████▏   | 1756/1845 [21:41<01:13,  1.21it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 95%|███████████████████████████████████████████████████████████████████████████▎   | 1759/1845 [21:43<01:04,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 96%|███████████████████████████████████████████████████████████████████████████▍   | 1762/1845 [21:45<01:04,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 96%|███████████████████████████████████████████████████████████████████████████▌   | 1765/1845 [21:47<01:00,  1.32it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 96%|███████████████████████████████████████████████████████████████████████████▋   | 1767/1845 [21:49<00:55,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 96%|███████████████████████████████████████████████████████████████████████████▊   | 1770/1845 [21:51<00:55,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 96%|███████████████████████████████████████████████████████████████████████████▉   | 1773/1845 [21:53<00:52,  1.36it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 96%|████████████████████████████████████████████████████████████████████████████   | 1776/1845 [21:55<00:47,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 96%|████████████████████████████████████████████████████████████████████████████▏  | 1779/1845 [21:57<00:46,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 97%|████████████████████████████████████████████████████████████████████████████▎  | 1781/1845 [21:59<00:51,  1.24it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 97%|████████████████████████████████████████████████████████████████████████████▍  | 1784/1845 [22:01<00:47,  1.28it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 97%|████████████████████████████████████████████████████████████████████████████▌  | 1787/1845 [22:03<00:45,  1.29it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 97%|████████████████████████████████████████████████████████████████████████████▌  | 1789/1845 [22:05<00:47,  1.19it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 97%|████████████████████████████████████████████████████████████████████████████▋  | 1792/1845 [22:07<00:37,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 97%|████████████████████████████████████████████████████████████████████████████▊  | 1795/1845 [22:09<00:35,  1.39it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 97%|████████████████████████████████████████████████████████████████████████████▉  | 1798/1845 [22:11<00:33,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 98%|█████████████████████████████████████████████████████████████████████████████  | 1801/1845 [22:13<00:30,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 98%|█████████████████████████████████████████████████████████████████████████████▏ | 1804/1845 [22:15<00:28,  1.42it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 98%|█████████████████████████████████████████████████████████████████████████████▎ | 1807/1845 [22:18<00:28,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 98%|█████████████████████████████████████████████████████████████████████████████▍ | 1809/1845 [22:19<00:26,  1.34it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 98%|█████████████████████████████████████████████████████████████████████████████▌ | 1812/1845 [22:21<00:24,  1.35it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 98%|█████████████████████████████████████████████████████████████████████████████▋ | 1815/1845 [22:24<00:22,  1.33it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 99%|█████████████████████████████████████████████████████████████████████████████▊ | 1818/1845 [22:26<00:19,  1.37it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 99%|█████████████████████████████████████████████████████████████████████████████▉ | 1820/1845 [22:27<00:19,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 99%|██████████████████████████████████████████████████████████████████████████████ | 1823/1845 [22:30<00:15,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 99%|██████████████████████████████████████████████████████████████████████████████▏| 1826/1845 [22:32<00:13,  1.41it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 99%|██████████████████████████████████████████████████████████████████████████████▎| 1829/1845 [22:34<00:12,  1.27it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 99%|██████████████████████████████████████████████████████████████████████████████▍| 1831/1845 [22:36<00:10,  1.31it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+ 99%|██████████████████████████████████████████████████████████████████████████████▌| 1835/1845 [22:38<00:06,  1.60it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+100%|██████████████████████████████████████████████████████████████████████████████▋| 1837/1845 [22:39<00:05,  1.47it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+100%|██████████████████████████████████████████████████████████████████████████████▊| 1840/1845 [22:42<00:03,  1.44it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+100%|██████████████████████████████████████████████████████████████████████████████▉| 1843/1845 [22:44<00:01,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+100%|███████████████████████████████████████████████████████████████████████████████| 1845/1845 [22:45<00:00,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+100%|███████████████████████████████████████████████████████████████████████████████| 1845/1845 [22:45<00:00,  1.40it/s]`use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Configuration saved in ./checkpoint-500/config.json                                                                     `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Feature extractor saved in ./checkpoint-500/preprocessor_config.json                                                    `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Feature extractor saved in ./checkpoint-500/preprocessor_config.json                                                    `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+Feature extractor saved in ./checkpoint-500/preprocessor_config.json                                                    `use_cache=True` is incompatible with gradient checkpointing. Setting `use_cache=False`...
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)
+05/04/2022 15:23:32 - WARNING - huggingface_hub.repository - Adding files tracked by Git LFS: ['wandb/run-20220504_142129-1tmxz74i/run-1tmxz74i.wandb']. This may take a bit of time if the files are large.
+huggingface/tokenizers: The current process just got forked, after parallelism has already been used. Disabling parallelism to avoid deadlocks...
+To disable this warning, you can either:
+	- Avoid using `tokenizers` before the fork if possible
+	- Explicitly set the environment variable TOKENIZERS_PARALLELISM=(true | false)