Upload training run run_20260408_135803

ParrotLLM end-of-training sync for local run runs/big_run/exp_c/run_20260408_135803.

.gitattributes

@@ -59,3 +59,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.mp4 filter=lfs diff=lfs merge=lfs -text
 *.webm filter=lfs diff=lfs merge=lfs -text
 runs/big_run/exp_c/run_20260408_124138/training_plots.pdf filter=lfs diff=lfs merge=lfs -text
+runs/big_run/exp_c/run_20260408_135803/training_plots.pdf filter=lfs diff=lfs merge=lfs -text

runs/big_run/exp_c/run_20260408_135803/.ipynb_checkpoints/config-checkpoint.json

@@ -0,0 +1,88 @@
+{
+  "preprocess": null,
+  "model": {
+    "vocab_size": 50258,
+    "pad_token_id": 50257,
+    "bos_token_id": 50256,
+    "eos_token_id": 50256,
+    "d_model": 384,
+    "n_layers": 14,
+    "n_heads": 6,
+    "d_ff": 768,
+    "context_length": 1024,
+    "bias": false,
+    "dropout": 0.005,
+    "rope_theta": 10000.0,
+    "gradient_checkpointing": false
+  },
+  "training": {
+    "device": "auto",
+    "train_bin": "data/exp_c/train.bin",
+    "val_bin": "data/exp_a/val.bin",
+    "num_workers": 4,
+    "pin_memory": true,
+    "batch_size": 32,
+    "gradient_accumulation_steps": 1,
+    "train_samples_per_epoch": null,
+    "val_sequence_stride": null,
+    "learning_rate": 0.001305,
+    "min_lr": 0.0001305,
+    "weight_decay": 0.063,
+    "beta1": 0.9,
+    "beta2": 0.9978,
+    "grad_clip": 1.786,
+    "warmup_steps": 305,
+    "max_steps": 3052,
+    "lr_schedule": "wsd",
+    "lr_decay_ratio": 0.179,
+    "z_loss_coeff": 0.000275,
+    "save_every": 250,
+    "eval_every": 100,
+    "checkpoint_dir": "checkpoints",
+    "keep_last_checkpoints": 10,
+    "keep_best_checkpoints": 10,
+    "early_stopping_patience": 0,
+    "early_stopping_min_delta": 0.001,
+    "runs_dir": "runs/big_run/exp_c",
+    "log_every": 10,
+    "hf_upload": {
+      "repo_id": "ParrotLabs/Preprocessed",
+      "repo_type": "dataset",
+      "path_in_repo": "",
+      "private": null
+    },
+    "compile": true
+  },
+  "tune": null,
+  "logging": {
+    "console_level": "INFO",
+    "file_level": "DEBUG",
+    "components": {
+      "data_preprocessing": "DEBUG",
+      "model_initialization": "INFO",
+      "training": "INFO"
+    },
+    "profiler": null
+  },
+  "eval": {
+    "device": "auto",
+    "batch_size": 32,
+    "max_sequences": 512,
+    "datasets": [
+      {
+        "name": "wikitext",
+        "path": "wikitext",
+        "subset": "wikitext-103-raw-v1",
+        "split": "test"
+      },
+      {
+        "name": "exp_c_val",
+        "path": "data/exp_c/val.bin",
+        "subset": null,
+        "split": null
+      }
+    ]
+  },
+  "inference": null,
+  "chat": null
+}

runs/big_run/exp_c/run_20260408_135803/.ipynb_checkpoints/train-checkpoint.log

@@ -0,0 +1,339 @@
+2026-04-08 13:58:04 - INFO - parrotllm - Logging initialised -> runs/big_run/exp_c/run_20260408_135803/train.log
+2026-04-08 13:58:04 - INFO - parrotllm.training - device=cuda:0 | rank=0 | world_size=8 | distributed=yes
+2026-04-08 13:58:04 - INFO - parrotllm.training - checkpoint directory=/home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints | keep_last=10 | keep_best=10
+2026-04-08 13:58:04 - INFO - parrotllm.training - train windows=random-overlapping | tokens=508793600 | valid_starts=508792576 | sampled_windows_per_epoch=496384
+2026-04-08 13:58:04 - INFO - parrotllm.training - val windows=sequential | tokens=8646900 | stride=1024 | windows=8444
+2026-04-08 13:58:04 - INFO - parrotllm.training - compiling model with torch.compile...
+2026-04-08 13:58:05 - INFO - parrotllm.training - 
+============================================================
+MODEL ARCHITECTURE SUMMARY
+============================================================
+
+Configuration:
+  Vocab size: 50258
+  Block size (context): 1024
+  Layers: 14
+  Heads: 6
+  Embedding dim: 384
+  FFN hidden dim: 768
+  Dropout: 0.005
+  Bias: False
+
+Parameters (unique, weight-tied layers counted once):
+  Total: 39,966,592 (39.97M)
+  Trainable: 39,966,592 (39.97M)
+  Non-trainable: 0
+  Non-embedding: 39,966,592 (39.97M)
+  Position embeddings: 0 (RoPE: 0 learned params)
+  Size (MB): 159.87
+
+Layer-wise breakdown (torchinfo):
+  Note: torchinfo double-counts weight-tied layers (tok_emb/lm_head).
+============================================================================================================================================
+Layer (type:depth-idx)                   Input Shape               Output Shape              Param #                   Trainable
+============================================================================================================================================
+ParrotLLM                                [32, 1024]                [32, 1024, 50258]         --                        True
+├─Embedding: 1-1                         [32, 1024]                [32, 1024, 384]           19,299,072                True
+├─Dropout: 1-2                           [32, 1024, 384]           [32, 1024, 384]           --                        --
+├─ModuleList: 1-3                        --                        --                        --                        True
+│    └─TransformerBlock: 2-1             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-1                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-2      [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-3                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-4                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-5               [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-6                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-2             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-7                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-8      [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-9                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-10                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-11              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-12                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-3             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-13                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-14     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-15                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-16                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-17              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    ���─RMSNorm: 3-18                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-4             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-19                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-20     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-21                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-22                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-23              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-24                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-5             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-25                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-26     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-27                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-28                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-29              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-30                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-6             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-31                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-32     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-33                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-34                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-35              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-36                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-7             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-37                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-38     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-39                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-40                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-41              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-42                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-8             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-43                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-44     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-45                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-46                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-47              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-48                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-9             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-49                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-50     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-51                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-52                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-53              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-54                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-10            [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-55                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-56     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-57                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-58                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-59              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-60                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-11            [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-61                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-62     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-63                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-64                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-65              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-66                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-12            [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-67                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-68     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-69                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-70                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-71              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-72                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-13            [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-73                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-74     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-75                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-76                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-77              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-78                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-14            [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-79                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-80     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-81                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-82                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-83              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-84                [32, 1024, 384]           [32, 1024, 384]           384                       True
+├─RMSNorm: 1-4                           [32, 1024, 384]           [32, 1024, 384]           384                       True
+├─Linear: 1-5                            [32, 1024, 384]           [32, 1024, 50258]         19,299,072                True
+============================================================================================================================================
+Total params: 59,265,664
+Trainable params: 59,265,664
+Non-trainable params: 0
+Total mult-adds (Units.GIGABYTES): 1.90
+============================================================================================================================================
+Input size (MB): 0.26
+Forward/backward pass size (MB): 34515.45
+Params size (MB): 237.06
+Estimated Total Size (MB): 34752.78
+============================================================================================================================================
+============================================================
+2026-04-08 13:58:05 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 13:58:14 - INFO - parrotllm.training -   Initial val: loss=10.8922, ppl=53754.44
+2026-04-08 13:58:30 - INFO - parrotllm.training - 
+============================================================
+Starting training...
+  Steps per epoch (approx): 1939
+  Max steps: 3052
+============================================================
+2026-04-08 13:58:30 - INFO - parrotllm.training - LR schedule=wsd | decay_ratio=0.179 | z_loss_coeff=0.000275
+2026-04-08 13:58:54 - INFO - parrotllm.training - step      1 | epoch 0 | loss 10.9251 | lr 4.28e-06 | grad 7.1913
+2026-04-08 13:58:57 - INFO - parrotllm.training - step     10 | epoch 0 | loss 10.2404 | lr 4.28e-05 | grad 1.8256
+2026-04-08 13:59:00 - INFO - parrotllm.training - step     20 | epoch 0 | loss 9.8834 | lr 8.56e-05 | grad 1.7080
+2026-04-08 13:59:03 - INFO - parrotllm.training - step     30 | epoch 0 | loss 9.3318 | lr 1.28e-04 | grad 1.6014
+2026-04-08 13:59:06 - INFO - parrotllm.training - step     40 | epoch 0 | loss 8.6693 | lr 1.71e-04 | grad 1.5937
+2026-04-08 13:59:09 - INFO - parrotllm.training - step     50 | epoch 0 | loss 8.0433 | lr 2.14e-04 | grad 1.3466
+2026-04-08 13:59:12 - INFO - parrotllm.training - step     60 | epoch 0 | loss 7.5934 | lr 2.57e-04 | grad 0.9873
+2026-04-08 13:59:15 - INFO - parrotllm.training - step     70 | epoch 0 | loss 7.3259 | lr 3.00e-04 | grad 0.6774
+2026-04-08 13:59:18 - INFO - parrotllm.training - step     80 | epoch 0 | loss 7.1066 | lr 3.42e-04 | grad 1.3694
+2026-04-08 13:59:21 - INFO - parrotllm.training - step     90 | epoch 0 | loss 6.8762 | lr 3.85e-04 | grad 0.5064
+2026-04-08 13:59:24 - INFO - parrotllm.training - step    100 | epoch 0 | loss 6.7364 | lr 4.28e-04 | grad 0.5288
+2026-04-08 13:59:24 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 13:59:24 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:00:00 - INFO - parrotllm.training -   Train: loss=6.7364, ppl=842.54
+2026-04-08 14:00:00 - INFO - parrotllm.training -   Val:   loss=6.7636, ppl=865.77
+2026-04-08 14:00:00 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:00:01 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_6p7636_epoch_0000_step_0000100.pt
+2026-04-08 14:00:02 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:00:05 - INFO - parrotllm.training - step    110 | epoch 0 | loss 6.6834 | lr 4.71e-04 | grad 1.5282
+2026-04-08 14:00:08 - INFO - parrotllm.training - step    120 | epoch 0 | loss 6.5337 | lr 5.13e-04 | grad 0.4935
+2026-04-08 14:00:11 - INFO - parrotllm.training - step    130 | epoch 0 | loss 6.3825 | lr 5.56e-04 | grad 0.3624
+2026-04-08 14:00:15 - INFO - parrotllm.training - step    140 | epoch 0 | loss 6.4176 | lr 5.99e-04 | grad 0.7881
+2026-04-08 14:00:18 - INFO - parrotllm.training - step    150 | epoch 0 | loss 6.3486 | lr 6.42e-04 | grad 0.3558
+2026-04-08 14:00:21 - INFO - parrotllm.training - step    160 | epoch 0 | loss 6.1151 | lr 6.85e-04 | grad 0.6742
+2026-04-08 14:00:24 - INFO - parrotllm.training - step    170 | epoch 0 | loss 6.1369 | lr 7.27e-04 | grad 0.4422
+2026-04-08 14:00:27 - INFO - parrotllm.training - step    180 | epoch 0 | loss 6.1238 | lr 7.70e-04 | grad 1.4895
+2026-04-08 14:00:30 - INFO - parrotllm.training - step    190 | epoch 0 | loss 6.0950 | lr 8.13e-04 | grad 0.4637
+2026-04-08 14:00:33 - INFO - parrotllm.training - step    200 | epoch 0 | loss 6.0127 | lr 8.56e-04 | grad 0.2755
+2026-04-08 14:00:33 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:00:33 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:00:39 - INFO - parrotllm.training -   Train: loss=6.0127, ppl=408.59
+2026-04-08 14:00:39 - INFO - parrotllm.training -   Val:   loss=5.9880, ppl=398.63
+2026-04-08 14:00:39 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:00:40 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_5p9880_epoch_0000_step_0000200.pt
+2026-04-08 14:00:40 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:00:43 - INFO - parrotllm.training - step    210 | epoch 0 | loss 5.9628 | lr 8.99e-04 | grad 0.6185
+2026-04-08 14:00:46 - INFO - parrotllm.training - step    220 | epoch 0 | loss 5.8730 | lr 9.41e-04 | grad 0.4188
+2026-04-08 14:00:50 - INFO - parrotllm.training - step    230 | epoch 0 | loss 5.8403 | lr 9.84e-04 | grad 0.8676
+2026-04-08 14:00:53 - INFO - parrotllm.training - step    240 | epoch 0 | loss 5.8026 | lr 1.03e-03 | grad 0.5417
+2026-04-08 14:00:56 - INFO - parrotllm.training - step    250 | epoch 0 | loss 5.7639 | lr 1.07e-03 | grad 0.5964
+2026-04-08 14:00:57 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0000250.pt
+2026-04-08 14:01:00 - INFO - parrotllm.training - step    260 | epoch 0 | loss 5.6420 | lr 1.11e-03 | grad 0.6763
+2026-04-08 14:01:03 - INFO - parrotllm.training - step    270 | epoch 0 | loss 5.6160 | lr 1.16e-03 | grad 1.0779
+2026-04-08 14:01:07 - INFO - parrotllm.training - step    280 | epoch 0 | loss 5.6631 | lr 1.20e-03 | grad 0.8367
+2026-04-08 14:01:10 - INFO - parrotllm.training - step    290 | epoch 0 | loss 5.5132 | lr 1.24e-03 | grad 0.3285
+2026-04-08 14:01:13 - INFO - parrotllm.training - step    300 | epoch 0 | loss 5.4597 | lr 1.28e-03 | grad 0.3686
+2026-04-08 14:01:13 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:01:13 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:01:18 - INFO - parrotllm.training -   Train: loss=5.4597, ppl=235.03
+2026-04-08 14:01:18 - INFO - parrotllm.training -   Val:   loss=5.4651, ppl=236.30
+2026-04-08 14:01:18 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:01:20 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_5p4651_epoch_0000_step_0000300.pt
+2026-04-08 14:01:20 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:01:23 - INFO - parrotllm.training - step    310 | epoch 0 | loss 5.4450 | lr 1.30e-03 | grad 0.6016
+2026-04-08 14:01:26 - INFO - parrotllm.training - step    320 | epoch 0 | loss 5.4612 | lr 1.30e-03 | grad 0.2626
+2026-04-08 14:01:29 - INFO - parrotllm.training - step    330 | epoch 0 | loss 5.3209 | lr 1.30e-03 | grad 0.3299
+2026-04-08 14:01:33 - INFO - parrotllm.training - step    340 | epoch 0 | loss 5.1924 | lr 1.30e-03 | grad 0.5871
+2026-04-08 14:01:36 - INFO - parrotllm.training - step    350 | epoch 0 | loss 5.2622 | lr 1.30e-03 | grad 0.4115
+2026-04-08 14:01:39 - INFO - parrotllm.training - step    360 | epoch 0 | loss 5.2611 | lr 1.30e-03 | grad 0.3492
+2026-04-08 14:01:42 - INFO - parrotllm.training - step    370 | epoch 0 | loss 5.3895 | lr 1.30e-03 | grad 0.6544
+2026-04-08 14:01:45 - INFO - parrotllm.training - step    380 | epoch 0 | loss 5.1282 | lr 1.30e-03 | grad 0.4954
+2026-04-08 14:01:48 - INFO - parrotllm.training - step    390 | epoch 0 | loss 5.0702 | lr 1.30e-03 | grad 0.4031
+2026-04-08 14:01:51 - INFO - parrotllm.training - step    400 | epoch 0 | loss 5.0852 | lr 1.30e-03 | grad 0.3539
+2026-04-08 14:01:51 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:01:51 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:01:57 - INFO - parrotllm.training -   Train: loss=5.0852, ppl=161.61
+2026-04-08 14:01:57 - INFO - parrotllm.training -   Val:   loss=5.0695, ppl=159.09
+2026-04-08 14:01:57 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:01:58 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_5p0695_epoch_0000_step_0000400.pt
+2026-04-08 14:01:58 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:02:01 - INFO - parrotllm.training - step    410 | epoch 0 | loss 5.0176 | lr 1.30e-03 | grad 0.3934
+2026-04-08 14:02:04 - INFO - parrotllm.training - step    420 | epoch 0 | loss 5.1627 | lr 1.30e-03 | grad 0.5370
+2026-04-08 14:02:08 - INFO - parrotllm.training - step    430 | epoch 0 | loss 4.9954 | lr 1.30e-03 | grad 0.4883
+2026-04-08 14:02:11 - INFO - parrotllm.training - step    440 | epoch 0 | loss 4.9951 | lr 1.30e-03 | grad 0.4548
+2026-04-08 14:02:14 - INFO - parrotllm.training - step    450 | epoch 0 | loss 4.9096 | lr 1.30e-03 | grad 0.3611
+2026-04-08 14:02:17 - INFO - parrotllm.training - step    460 | epoch 0 | loss 4.8739 | lr 1.30e-03 | grad 0.3378
+2026-04-08 14:02:20 - INFO - parrotllm.training - step    470 | epoch 0 | loss 4.8027 | lr 1.30e-03 | grad 0.3606
+2026-04-08 14:02:23 - INFO - parrotllm.training - step    480 | epoch 0 | loss 4.8119 | lr 1.30e-03 | grad 0.4046
+2026-04-08 14:02:26 - INFO - parrotllm.training - step    490 | epoch 0 | loss 4.9417 | lr 1.30e-03 | grad 0.6448
+2026-04-08 14:02:29 - INFO - parrotllm.training - step    500 | epoch 0 | loss 4.7433 | lr 1.30e-03 | grad 0.3431
+2026-04-08 14:02:29 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:02:29 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:02:35 - INFO - parrotllm.training -   Train: loss=4.7433, ppl=114.81
+2026-04-08 14:02:35 - INFO - parrotllm.training -   Val:   loss=4.7606, ppl=116.82
+2026-04-08 14:02:35 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:02:36 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p7606_epoch_0000_step_0000500.pt
+2026-04-08 14:02:37 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:02:38 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0000500.pt
+2026-04-08 14:02:41 - INFO - parrotllm.training - step    510 | epoch 0 | loss 4.6827 | lr 1.30e-03 | grad 0.4709
+2026-04-08 14:02:45 - INFO - parrotllm.training - step    520 | epoch 0 | loss 4.8279 | lr 1.30e-03 | grad 0.4272
+2026-04-08 14:02:48 - INFO - parrotllm.training - step    530 | epoch 0 | loss 4.7086 | lr 1.30e-03 | grad 0.3580
+2026-04-08 14:02:51 - INFO - parrotllm.training - step    540 | epoch 0 | loss 4.7267 | lr 1.30e-03 | grad 0.3677
+2026-04-08 14:02:54 - INFO - parrotllm.training - step    550 | epoch 0 | loss 4.6257 | lr 1.30e-03 | grad 0.5231
+2026-04-08 14:02:57 - INFO - parrotllm.training - step    560 | epoch 0 | loss 4.5377 | lr 1.30e-03 | grad 0.3239
+2026-04-08 14:03:00 - INFO - parrotllm.training - step    570 | epoch 0 | loss 4.5855 | lr 1.30e-03 | grad 0.3071
+2026-04-08 14:03:04 - INFO - parrotllm.training - step    580 | epoch 0 | loss 4.3974 | lr 1.30e-03 | grad 0.3457
+2026-04-08 14:03:07 - INFO - parrotllm.training - step    590 | epoch 0 | loss 4.5168 | lr 1.30e-03 | grad 0.5359
+2026-04-08 14:03:10 - INFO - parrotllm.training - step    600 | epoch 0 | loss 4.5400 | lr 1.30e-03 | grad 0.3343
+2026-04-08 14:03:10 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:03:10 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:03:16 - INFO - parrotllm.training -   Train: loss=4.5400, ppl=93.69
+2026-04-08 14:03:16 - INFO - parrotllm.training -   Val:   loss=4.5068, ppl=90.63
+2026-04-08 14:03:16 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:03:17 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p5068_epoch_0000_step_0000600.pt
+2026-04-08 14:03:17 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:03:20 - INFO - parrotllm.training - step    610 | epoch 0 | loss 4.4753 | lr 1.30e-03 | grad 0.2855
+2026-04-08 14:03:24 - INFO - parrotllm.training - step    620 | epoch 0 | loss 4.4947 | lr 1.30e-03 | grad 0.3025
+2026-04-08 14:03:27 - INFO - parrotllm.training - step    630 | epoch 0 | loss 4.4394 | lr 1.30e-03 | grad 0.4398
+2026-04-08 14:03:30 - INFO - parrotllm.training - step    640 | epoch 0 | loss 4.4087 | lr 1.30e-03 | grad 0.2784
+2026-04-08 14:03:33 - INFO - parrotllm.training - step    650 | epoch 0 | loss 4.4705 | lr 1.30e-03 | grad 0.2882
+2026-04-08 14:03:36 - INFO - parrotllm.training - step    660 | epoch 0 | loss 4.3001 | lr 1.30e-03 | grad 0.4116
+2026-04-08 14:03:39 - INFO - parrotllm.training - step    670 | epoch 0 | loss 4.4093 | lr 1.30e-03 | grad 0.3227
+2026-04-08 14:03:42 - INFO - parrotllm.training - step    680 | epoch 0 | loss 4.4190 | lr 1.30e-03 | grad 0.3476
+2026-04-08 14:03:45 - INFO - parrotllm.training - step    690 | epoch 0 | loss 4.2442 | lr 1.30e-03 | grad 0.3783
+2026-04-08 14:03:48 - INFO - parrotllm.training - step    700 | epoch 0 | loss 4.3616 | lr 1.30e-03 | grad 0.5368
+2026-04-08 14:03:48 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:03:48 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:03:54 - INFO - parrotllm.training -   Train: loss=4.3616, ppl=78.39
+2026-04-08 14:03:54 - INFO - parrotllm.training -   Val:   loss=4.3528, ppl=77.70
+2026-04-08 14:03:54 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:03:55 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p3528_epoch_0000_step_0000700.pt
+2026-04-08 14:03:56 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:03:59 - INFO - parrotllm.training - step    710 | epoch 0 | loss 4.2724 | lr 1.30e-03 | grad 0.2602
+2026-04-08 14:04:02 - INFO - parrotllm.training - step    720 | epoch 0 | loss 4.3498 | lr 1.30e-03 | grad 0.3388
+2026-04-08 14:04:05 - INFO - parrotllm.training - step    730 | epoch 0 | loss 4.2156 | lr 1.30e-03 | grad 0.3329
+2026-04-08 14:04:08 - INFO - parrotllm.training - step    740 | epoch 0 | loss 4.2054 | lr 1.30e-03 | grad 0.2879
+2026-04-08 14:04:11 - INFO - parrotllm.training - step    750 | epoch 0 | loss 4.2538 | lr 1.30e-03 | grad 0.3625
+2026-04-08 14:04:12 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0000750.pt
+2026-04-08 14:04:16 - INFO - parrotllm.training - step    760 | epoch 0 | loss 4.2472 | lr 1.30e-03 | grad 0.2917
+2026-04-08 14:04:19 - INFO - parrotllm.training - step    770 | epoch 0 | loss 4.1726 | lr 1.30e-03 | grad 0.3386
+2026-04-08 14:04:22 - INFO - parrotllm.training - step    780 | epoch 0 | loss 4.2424 | lr 1.30e-03 | grad 0.2958
+2026-04-08 14:04:25 - INFO - parrotllm.training - step    790 | epoch 0 | loss 4.2461 | lr 1.30e-03 | grad 0.3299
+2026-04-08 14:04:28 - INFO - parrotllm.training - step    800 | epoch 0 | loss 4.2986 | lr 1.30e-03 | grad 0.3326
+2026-04-08 14:04:28 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:04:28 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:04:34 - INFO - parrotllm.training -   Train: loss=4.2986, ppl=73.60
+2026-04-08 14:04:34 - INFO - parrotllm.training -   Val:   loss=4.2289, ppl=68.64
+2026-04-08 14:04:34 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:04:35 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p2289_epoch_0000_step_0000800.pt
+2026-04-08 14:04:36 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:04:39 - INFO - parrotllm.training - step    810 | epoch 0 | loss 4.1100 | lr 1.30e-03 | grad 0.3297
+2026-04-08 14:04:42 - INFO - parrotllm.training - step    820 | epoch 0 | loss 4.1684 | lr 1.30e-03 | grad 0.4951
+2026-04-08 14:04:45 - INFO - parrotllm.training - step    830 | epoch 0 | loss 4.2305 | lr 1.30e-03 | grad 0.2999
+2026-04-08 14:04:48 - INFO - parrotllm.training - step    840 | epoch 0 | loss 4.1809 | lr 1.30e-03 | grad 0.2789
+2026-04-08 14:04:51 - INFO - parrotllm.training - step    850 | epoch 0 | loss 4.1167 | lr 1.30e-03 | grad 0.3522
+2026-04-08 14:04:54 - INFO - parrotllm.training - step    860 | epoch 0 | loss 4.0878 | lr 1.30e-03 | grad 0.3210
+2026-04-08 14:04:57 - INFO - parrotllm.training - step    870 | epoch 0 | loss 4.1254 | lr 1.30e-03 | grad 0.3781
+2026-04-08 14:05:01 - INFO - parrotllm.training - step    880 | epoch 0 | loss 4.2074 | lr 1.30e-03 | grad 0.3715
+2026-04-08 14:05:04 - INFO - parrotllm.training - step    890 | epoch 0 | loss 4.1691 | lr 1.30e-03 | grad 0.3186
+2026-04-08 14:05:07 - INFO - parrotllm.training - step    900 | epoch 0 | loss 4.1067 | lr 1.30e-03 | grad 0.3043
+2026-04-08 14:05:07 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:05:07 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:05:13 - INFO - parrotllm.training -   Train: loss=4.1067, ppl=60.75
+2026-04-08 14:05:13 - INFO - parrotllm.training -   Val:   loss=4.1455, ppl=63.15
+2026-04-08 14:05:13 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:05:14 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p1455_epoch_0000_step_0000900.pt
+2026-04-08 14:05:14 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:05:17 - INFO - parrotllm.training - step    910 | epoch 0 | loss 4.1516 | lr 1.30e-03 | grad 0.3303
+2026-04-08 14:05:20 - INFO - parrotllm.training - step    920 | epoch 0 | loss 4.1988 | lr 1.30e-03 | grad 0.2828
+2026-04-08 14:05:23 - INFO - parrotllm.training - step    930 | epoch 0 | loss 4.1395 | lr 1.30e-03 | grad 0.2683
+2026-04-08 14:05:27 - INFO - parrotllm.training - step    940 | epoch 0 | loss 4.1705 | lr 1.30e-03 | grad 0.2802
+2026-04-08 14:05:30 - INFO - parrotllm.training - step    950 | epoch 0 | loss 4.1899 | lr 1.30e-03 | grad 0.2848
+2026-04-08 14:05:33 - INFO - parrotllm.training - step    960 | epoch 0 | loss 4.0292 | lr 1.30e-03 | grad 0.3756
+2026-04-08 14:05:36 - INFO - parrotllm.training - step    970 | epoch 0 | loss 4.2175 | lr 1.30e-03 | grad 0.3247
+2026-04-08 14:05:39 - INFO - parrotllm.training - step    980 | epoch 0 | loss 4.0566 | lr 1.30e-03 | grad 0.2580
+2026-04-08 14:05:42 - INFO - parrotllm.training - step    990 | epoch 0 | loss 3.9257 | lr 1.30e-03 | grad 0.3235
+2026-04-08 14:05:45 - INFO - parrotllm.training - step   1000 | epoch 0 | loss 4.1197 | lr 1.30e-03 | grad 0.3498
+2026-04-08 14:05:45 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:05:45 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:05:51 - INFO - parrotllm.training -   Train: loss=4.1197, ppl=61.54
+2026-04-08 14:05:51 - INFO - parrotllm.training -   Val:   loss=4.0852, ppl=59.46
+2026-04-08 14:05:51 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:05:52 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p0852_epoch_0000_step_0001000.pt
+2026-04-08 14:05:52 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:05:54 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0001000.pt
+2026-04-08 14:05:57 - INFO - parrotllm.training - step   1010 | epoch 0 | loss 3.9842 | lr 1.30e-03 | grad 0.2686
+2026-04-08 14:06:01 - INFO - parrotllm.training - step   1020 | epoch 0 | loss 4.0029 | lr 1.30e-03 | grad 0.3144
+2026-04-08 14:06:04 - INFO - parrotllm.training - step   1030 | epoch 0 | loss 4.0865 | lr 1.30e-03 | grad 0.3201
+2026-04-08 14:06:07 - INFO - parrotllm.training - step   1040 | epoch 0 | loss 4.0908 | lr 1.30e-03 | grad 0.2985
+2026-04-08 14:06:10 - INFO - parrotllm.training - step   1050 | epoch 0 | loss 4.0157 | lr 1.30e-03 | grad 0.2554

runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p6338_epoch_0001_step_0003000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f85d78532dc48d951a1bfe25a0836a9aa778c4eb33c0c63bc4907c0fe9c1afd
+size 479869500

runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p6558_epoch_0001_step_0002900.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b4a284d0716ffdc038466f8e21365706e7d1a17808a8087f01b6f129a77b6900
+size 479869500

runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p6796_epoch_0001_step_0002800.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f58f5282328f46362bcf89d59ad87c36cfe4ec1e0a4f46b07621ac44c2c9f780
+size 479869500

runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7026_epoch_0001_step_0002700.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e031a2765edec85cc94dd28cd4578224e1fa2a7b27b10f3e6d72b358e52bd8a
+size 479869500

runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7307_epoch_0001_step_0002600.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22e9285a902475fb8ef5f9d61c931bc275a1d5f67df2b6acaa6d2f466c7a3cd6
+size 479869500

runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7530_epoch_0001_step_0002500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:551c6078725bfe56d90ff06e4edb3d856bd9ac05cd352a797156b014779b428a
+size 479869500

runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7614_epoch_0001_step_0002400.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58ec207f6544c087010533f384178b91d5661ba26c5fc2c35255dde68c8b7312
+size 479869500

runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7749_epoch_0001_step_0002300.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:55a997370eb03bf71ce28f6465c5d646eb0528951b408e96f8836bcc69778c47
+size 479869500

runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7865_epoch_0001_step_0002200.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f4e84ed82677f4611cb37cf7a2293055c92d5072aab3c14ed1a94f8c568d70b4
+size 479869500

runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7972_epoch_0001_step_0002100.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e2b604c398481655693b543ba468c6d08de6bcc49893ac569f3cc366279c3308
+size 479869500

runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0001000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:841ca156ccdfcc40367017e70fa63812710ab162ab7354ab75f3057c0652b2c9
+size 479860488

runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0001250.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8d609ce52e0d085aef8d5f2e0cb300cb81631dd2066d89945dd119283c47a2c4
+size 479860488

runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0001500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c5d2c0b400eee2bc4757a3bda89ee9ca91f86ff2925d8281ddf411f96f0c7f74
+size 479860488

runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0001750.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b11eac4a3dffa29b21cf1edafe2fb61b22eb5d7886741d0fade55cfdb3721a15
+size 479860488

runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0002000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:aa22dfd1e151fa7364517d8331c041bc1868e220e3b2d7d668ab31421bc65dbf
+size 479860488

runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0002250.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:def05f0c7ca05694db63a3b6abb853e8abcf3f15a3d2dd0010b3b5bff4f8b9ca
+size 479860488

runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0002500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:94bb3754cb0777a17de34af590cdee3db52fab943331bae133ce4f171ade13b3
+size 479860488

runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0002750.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a3396714b705ffad3135276900f110782d6534b32724ff9ecc0547f36d6227bd
+size 479860488

runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0003000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68e57fdbc786eef53bd2fcc2a070d5f15391cb29e9b4273edb316da09725fc91
+size 479860488

runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0003052.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9c4b81de0041acc27b5463bac30bf88b89656d0b40cd348702ba5376c9acd0ea
+size 479860488

runs/big_run/exp_c/run_20260408_135803/config.json

@@ -0,0 +1,88 @@
+{
+  "preprocess": null,
+  "model": {
+    "vocab_size": 50258,
+    "pad_token_id": 50257,
+    "bos_token_id": 50256,
+    "eos_token_id": 50256,
+    "d_model": 384,
+    "n_layers": 14,
+    "n_heads": 6,
+    "d_ff": 768,
+    "context_length": 1024,
+    "bias": false,
+    "dropout": 0.005,
+    "rope_theta": 10000.0,
+    "gradient_checkpointing": false
+  },
+  "training": {
+    "device": "auto",
+    "train_bin": "data/exp_c/train.bin",
+    "val_bin": "data/exp_a/val.bin",
+    "num_workers": 4,
+    "pin_memory": true,
+    "batch_size": 32,
+    "gradient_accumulation_steps": 1,
+    "train_samples_per_epoch": null,
+    "val_sequence_stride": null,
+    "learning_rate": 0.001305,
+    "min_lr": 0.0001305,
+    "weight_decay": 0.063,
+    "beta1": 0.9,
+    "beta2": 0.9978,
+    "grad_clip": 1.786,
+    "warmup_steps": 305,
+    "max_steps": 3052,
+    "lr_schedule": "wsd",
+    "lr_decay_ratio": 0.179,
+    "z_loss_coeff": 0.000275,
+    "save_every": 250,
+    "eval_every": 100,
+    "checkpoint_dir": "checkpoints",
+    "keep_last_checkpoints": 10,
+    "keep_best_checkpoints": 10,
+    "early_stopping_patience": 0,
+    "early_stopping_min_delta": 0.001,
+    "runs_dir": "runs/big_run/exp_c",
+    "log_every": 10,
+    "hf_upload": {
+      "repo_id": "ParrotLabs/Preprocessed",
+      "repo_type": "dataset",
+      "path_in_repo": "",
+      "private": null
+    },
+    "compile": true
+  },
+  "tune": null,
+  "logging": {
+    "console_level": "INFO",
+    "file_level": "DEBUG",
+    "components": {
+      "data_preprocessing": "DEBUG",
+      "model_initialization": "INFO",
+      "training": "INFO"
+    },
+    "profiler": null
+  },
+  "eval": {
+    "device": "auto",
+    "batch_size": 32,
+    "max_sequences": 512,
+    "datasets": [
+      {
+        "name": "wikitext",
+        "path": "wikitext",
+        "subset": "wikitext-103-raw-v1",
+        "split": "test"
+      },
+      {
+        "name": "exp_c_val",
+        "path": "data/exp_c/val.bin",
+        "subset": null,
+        "split": null
+      }
+    ]
+  },
+  "inference": null,
+  "chat": null
+}

runs/big_run/exp_c/run_20260408_135803/train.log

@@ -0,0 +1,721 @@
+2026-04-08 13:58:04 - INFO - parrotllm - Logging initialised -> runs/big_run/exp_c/run_20260408_135803/train.log
+2026-04-08 13:58:04 - INFO - parrotllm.training - device=cuda:0 | rank=0 | world_size=8 | distributed=yes
+2026-04-08 13:58:04 - INFO - parrotllm.training - checkpoint directory=/home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints | keep_last=10 | keep_best=10
+2026-04-08 13:58:04 - INFO - parrotllm.training - train windows=random-overlapping | tokens=508793600 | valid_starts=508792576 | sampled_windows_per_epoch=496384
+2026-04-08 13:58:04 - INFO - parrotllm.training - val windows=sequential | tokens=8646900 | stride=1024 | windows=8444
+2026-04-08 13:58:04 - INFO - parrotllm.training - compiling model with torch.compile...
+2026-04-08 13:58:05 - INFO - parrotllm.training - 
+============================================================
+MODEL ARCHITECTURE SUMMARY
+============================================================
+
+Configuration:
+  Vocab size: 50258
+  Block size (context): 1024
+  Layers: 14
+  Heads: 6
+  Embedding dim: 384
+  FFN hidden dim: 768
+  Dropout: 0.005
+  Bias: False
+
+Parameters (unique, weight-tied layers counted once):
+  Total: 39,966,592 (39.97M)
+  Trainable: 39,966,592 (39.97M)
+  Non-trainable: 0
+  Non-embedding: 39,966,592 (39.97M)
+  Position embeddings: 0 (RoPE: 0 learned params)
+  Size (MB): 159.87
+
+Layer-wise breakdown (torchinfo):
+  Note: torchinfo double-counts weight-tied layers (tok_emb/lm_head).
+============================================================================================================================================
+Layer (type:depth-idx)                   Input Shape               Output Shape              Param #                   Trainable
+============================================================================================================================================
+ParrotLLM                                [32, 1024]                [32, 1024, 50258]         --                        True
+├─Embedding: 1-1                         [32, 1024]                [32, 1024, 384]           19,299,072                True
+├─Dropout: 1-2                           [32, 1024, 384]           [32, 1024, 384]           --                        --
+├─ModuleList: 1-3                        --                        --                        --                        True
+│    └─TransformerBlock: 2-1             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-1                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-2      [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-3                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-4                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-5               [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-6                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-2             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-7                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-8      [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-9                 [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-10                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-11              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-12                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-3             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-13                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-14     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-15                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-16                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-17              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    ���─RMSNorm: 3-18                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-4             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-19                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-20     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-21                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-22                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-23              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-24                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-5             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-25                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-26     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-27                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-28                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-29              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-30                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-6             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-31                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-32     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-33                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-34                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-35              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-36                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-7             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-37                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-38     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-39                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-40                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-41              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-42                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-8             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-43                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-44     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-45                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-46                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-47              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-48                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-9             [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-49                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-50     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-51                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-52                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-53              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-54                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-10            [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-55                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-56     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-57                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-58                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-59              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-60                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-11            [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-61                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-62     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-63                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-64                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-65              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-66                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-12            [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-67                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-68     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-69                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-70                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-71              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-72                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-13            [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-73                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-74     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-75                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-76                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-77              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-78                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    └─TransformerBlock: 2-14            [32, 1024, 384]           [32, 1024, 384]           --                        True
+│    │    └─RMSNorm: 3-79                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─MultiHeadAttention: 3-80     [32, 1024, 384]           [32, 1024, 384]           589,952                   True
+│    │    └─RMSNorm: 3-81                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─RMSNorm: 3-82                [32, 1024, 384]           [32, 1024, 384]           384                       True
+│    │    └─SwiGLUMLP: 3-83              [32, 1024, 384]           [32, 1024, 384]           884,736                   True
+│    │    └─RMSNorm: 3-84                [32, 1024, 384]           [32, 1024, 384]           384                       True
+├─RMSNorm: 1-4                           [32, 1024, 384]           [32, 1024, 384]           384                       True
+├─Linear: 1-5                            [32, 1024, 384]           [32, 1024, 50258]         19,299,072                True
+============================================================================================================================================
+Total params: 59,265,664
+Trainable params: 59,265,664
+Non-trainable params: 0
+Total mult-adds (Units.GIGABYTES): 1.90
+============================================================================================================================================
+Input size (MB): 0.26
+Forward/backward pass size (MB): 34515.45
+Params size (MB): 237.06
+Estimated Total Size (MB): 34752.78
+============================================================================================================================================
+============================================================
+2026-04-08 13:58:05 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 13:58:14 - INFO - parrotllm.training -   Initial val: loss=10.8922, ppl=53754.44
+2026-04-08 13:58:30 - INFO - parrotllm.training - 
+============================================================
+Starting training...
+  Steps per epoch (approx): 1939
+  Max steps: 3052
+============================================================
+2026-04-08 13:58:30 - INFO - parrotllm.training - LR schedule=wsd | decay_ratio=0.179 | z_loss_coeff=0.000275
+2026-04-08 13:58:54 - INFO - parrotllm.training - step      1 | epoch 0 | loss 10.9251 | lr 4.28e-06 | grad 7.1913
+2026-04-08 13:58:57 - INFO - parrotllm.training - step     10 | epoch 0 | loss 10.2404 | lr 4.28e-05 | grad 1.8256
+2026-04-08 13:59:00 - INFO - parrotllm.training - step     20 | epoch 0 | loss 9.8834 | lr 8.56e-05 | grad 1.7080
+2026-04-08 13:59:03 - INFO - parrotllm.training - step     30 | epoch 0 | loss 9.3318 | lr 1.28e-04 | grad 1.6014
+2026-04-08 13:59:06 - INFO - parrotllm.training - step     40 | epoch 0 | loss 8.6693 | lr 1.71e-04 | grad 1.5937
+2026-04-08 13:59:09 - INFO - parrotllm.training - step     50 | epoch 0 | loss 8.0433 | lr 2.14e-04 | grad 1.3466
+2026-04-08 13:59:12 - INFO - parrotllm.training - step     60 | epoch 0 | loss 7.5934 | lr 2.57e-04 | grad 0.9873
+2026-04-08 13:59:15 - INFO - parrotllm.training - step     70 | epoch 0 | loss 7.3259 | lr 3.00e-04 | grad 0.6774
+2026-04-08 13:59:18 - INFO - parrotllm.training - step     80 | epoch 0 | loss 7.1066 | lr 3.42e-04 | grad 1.3694
+2026-04-08 13:59:21 - INFO - parrotllm.training - step     90 | epoch 0 | loss 6.8762 | lr 3.85e-04 | grad 0.5064
+2026-04-08 13:59:24 - INFO - parrotllm.training - step    100 | epoch 0 | loss 6.7364 | lr 4.28e-04 | grad 0.5288
+2026-04-08 13:59:24 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 13:59:24 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:00:00 - INFO - parrotllm.training -   Train: loss=6.7364, ppl=842.54
+2026-04-08 14:00:00 - INFO - parrotllm.training -   Val:   loss=6.7636, ppl=865.77
+2026-04-08 14:00:00 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:00:01 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_6p7636_epoch_0000_step_0000100.pt
+2026-04-08 14:00:02 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:00:05 - INFO - parrotllm.training - step    110 | epoch 0 | loss 6.6834 | lr 4.71e-04 | grad 1.5282
+2026-04-08 14:00:08 - INFO - parrotllm.training - step    120 | epoch 0 | loss 6.5337 | lr 5.13e-04 | grad 0.4935
+2026-04-08 14:00:11 - INFO - parrotllm.training - step    130 | epoch 0 | loss 6.3825 | lr 5.56e-04 | grad 0.3624
+2026-04-08 14:00:15 - INFO - parrotllm.training - step    140 | epoch 0 | loss 6.4176 | lr 5.99e-04 | grad 0.7881
+2026-04-08 14:00:18 - INFO - parrotllm.training - step    150 | epoch 0 | loss 6.3486 | lr 6.42e-04 | grad 0.3558
+2026-04-08 14:00:21 - INFO - parrotllm.training - step    160 | epoch 0 | loss 6.1151 | lr 6.85e-04 | grad 0.6742
+2026-04-08 14:00:24 - INFO - parrotllm.training - step    170 | epoch 0 | loss 6.1369 | lr 7.27e-04 | grad 0.4422
+2026-04-08 14:00:27 - INFO - parrotllm.training - step    180 | epoch 0 | loss 6.1238 | lr 7.70e-04 | grad 1.4895
+2026-04-08 14:00:30 - INFO - parrotllm.training - step    190 | epoch 0 | loss 6.0950 | lr 8.13e-04 | grad 0.4637
+2026-04-08 14:00:33 - INFO - parrotllm.training - step    200 | epoch 0 | loss 6.0127 | lr 8.56e-04 | grad 0.2755
+2026-04-08 14:00:33 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:00:33 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:00:39 - INFO - parrotllm.training -   Train: loss=6.0127, ppl=408.59
+2026-04-08 14:00:39 - INFO - parrotllm.training -   Val:   loss=5.9880, ppl=398.63
+2026-04-08 14:00:39 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:00:40 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_5p9880_epoch_0000_step_0000200.pt
+2026-04-08 14:00:40 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:00:43 - INFO - parrotllm.training - step    210 | epoch 0 | loss 5.9628 | lr 8.99e-04 | grad 0.6185
+2026-04-08 14:00:46 - INFO - parrotllm.training - step    220 | epoch 0 | loss 5.8730 | lr 9.41e-04 | grad 0.4188
+2026-04-08 14:00:50 - INFO - parrotllm.training - step    230 | epoch 0 | loss 5.8403 | lr 9.84e-04 | grad 0.8676
+2026-04-08 14:00:53 - INFO - parrotllm.training - step    240 | epoch 0 | loss 5.8026 | lr 1.03e-03 | grad 0.5417
+2026-04-08 14:00:56 - INFO - parrotllm.training - step    250 | epoch 0 | loss 5.7639 | lr 1.07e-03 | grad 0.5964
+2026-04-08 14:00:57 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0000250.pt
+2026-04-08 14:01:00 - INFO - parrotllm.training - step    260 | epoch 0 | loss 5.6420 | lr 1.11e-03 | grad 0.6763
+2026-04-08 14:01:03 - INFO - parrotllm.training - step    270 | epoch 0 | loss 5.6160 | lr 1.16e-03 | grad 1.0779
+2026-04-08 14:01:07 - INFO - parrotllm.training - step    280 | epoch 0 | loss 5.6631 | lr 1.20e-03 | grad 0.8367
+2026-04-08 14:01:10 - INFO - parrotllm.training - step    290 | epoch 0 | loss 5.5132 | lr 1.24e-03 | grad 0.3285
+2026-04-08 14:01:13 - INFO - parrotllm.training - step    300 | epoch 0 | loss 5.4597 | lr 1.28e-03 | grad 0.3686
+2026-04-08 14:01:13 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:01:13 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:01:18 - INFO - parrotllm.training -   Train: loss=5.4597, ppl=235.03
+2026-04-08 14:01:18 - INFO - parrotllm.training -   Val:   loss=5.4651, ppl=236.30
+2026-04-08 14:01:18 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:01:20 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_5p4651_epoch_0000_step_0000300.pt
+2026-04-08 14:01:20 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:01:23 - INFO - parrotllm.training - step    310 | epoch 0 | loss 5.4450 | lr 1.30e-03 | grad 0.6016
+2026-04-08 14:01:26 - INFO - parrotllm.training - step    320 | epoch 0 | loss 5.4612 | lr 1.30e-03 | grad 0.2626
+2026-04-08 14:01:29 - INFO - parrotllm.training - step    330 | epoch 0 | loss 5.3209 | lr 1.30e-03 | grad 0.3299
+2026-04-08 14:01:33 - INFO - parrotllm.training - step    340 | epoch 0 | loss 5.1924 | lr 1.30e-03 | grad 0.5871
+2026-04-08 14:01:36 - INFO - parrotllm.training - step    350 | epoch 0 | loss 5.2622 | lr 1.30e-03 | grad 0.4115
+2026-04-08 14:01:39 - INFO - parrotllm.training - step    360 | epoch 0 | loss 5.2611 | lr 1.30e-03 | grad 0.3492
+2026-04-08 14:01:42 - INFO - parrotllm.training - step    370 | epoch 0 | loss 5.3895 | lr 1.30e-03 | grad 0.6544
+2026-04-08 14:01:45 - INFO - parrotllm.training - step    380 | epoch 0 | loss 5.1282 | lr 1.30e-03 | grad 0.4954
+2026-04-08 14:01:48 - INFO - parrotllm.training - step    390 | epoch 0 | loss 5.0702 | lr 1.30e-03 | grad 0.4031
+2026-04-08 14:01:51 - INFO - parrotllm.training - step    400 | epoch 0 | loss 5.0852 | lr 1.30e-03 | grad 0.3539
+2026-04-08 14:01:51 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:01:51 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:01:57 - INFO - parrotllm.training -   Train: loss=5.0852, ppl=161.61
+2026-04-08 14:01:57 - INFO - parrotllm.training -   Val:   loss=5.0695, ppl=159.09
+2026-04-08 14:01:57 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:01:58 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_5p0695_epoch_0000_step_0000400.pt
+2026-04-08 14:01:58 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:02:01 - INFO - parrotllm.training - step    410 | epoch 0 | loss 5.0176 | lr 1.30e-03 | grad 0.3934
+2026-04-08 14:02:04 - INFO - parrotllm.training - step    420 | epoch 0 | loss 5.1627 | lr 1.30e-03 | grad 0.5370
+2026-04-08 14:02:08 - INFO - parrotllm.training - step    430 | epoch 0 | loss 4.9954 | lr 1.30e-03 | grad 0.4883
+2026-04-08 14:02:11 - INFO - parrotllm.training - step    440 | epoch 0 | loss 4.9951 | lr 1.30e-03 | grad 0.4548
+2026-04-08 14:02:14 - INFO - parrotllm.training - step    450 | epoch 0 | loss 4.9096 | lr 1.30e-03 | grad 0.3611
+2026-04-08 14:02:17 - INFO - parrotllm.training - step    460 | epoch 0 | loss 4.8739 | lr 1.30e-03 | grad 0.3378
+2026-04-08 14:02:20 - INFO - parrotllm.training - step    470 | epoch 0 | loss 4.8027 | lr 1.30e-03 | grad 0.3606
+2026-04-08 14:02:23 - INFO - parrotllm.training - step    480 | epoch 0 | loss 4.8119 | lr 1.30e-03 | grad 0.4046
+2026-04-08 14:02:26 - INFO - parrotllm.training - step    490 | epoch 0 | loss 4.9417 | lr 1.30e-03 | grad 0.6448
+2026-04-08 14:02:29 - INFO - parrotllm.training - step    500 | epoch 0 | loss 4.7433 | lr 1.30e-03 | grad 0.3431
+2026-04-08 14:02:29 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:02:29 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:02:35 - INFO - parrotllm.training -   Train: loss=4.7433, ppl=114.81
+2026-04-08 14:02:35 - INFO - parrotllm.training -   Val:   loss=4.7606, ppl=116.82
+2026-04-08 14:02:35 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:02:36 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p7606_epoch_0000_step_0000500.pt
+2026-04-08 14:02:37 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:02:38 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0000500.pt
+2026-04-08 14:02:41 - INFO - parrotllm.training - step    510 | epoch 0 | loss 4.6827 | lr 1.30e-03 | grad 0.4709
+2026-04-08 14:02:45 - INFO - parrotllm.training - step    520 | epoch 0 | loss 4.8279 | lr 1.30e-03 | grad 0.4272
+2026-04-08 14:02:48 - INFO - parrotllm.training - step    530 | epoch 0 | loss 4.7086 | lr 1.30e-03 | grad 0.3580
+2026-04-08 14:02:51 - INFO - parrotllm.training - step    540 | epoch 0 | loss 4.7267 | lr 1.30e-03 | grad 0.3677
+2026-04-08 14:02:54 - INFO - parrotllm.training - step    550 | epoch 0 | loss 4.6257 | lr 1.30e-03 | grad 0.5231
+2026-04-08 14:02:57 - INFO - parrotllm.training - step    560 | epoch 0 | loss 4.5377 | lr 1.30e-03 | grad 0.3239
+2026-04-08 14:03:00 - INFO - parrotllm.training - step    570 | epoch 0 | loss 4.5855 | lr 1.30e-03 | grad 0.3071
+2026-04-08 14:03:04 - INFO - parrotllm.training - step    580 | epoch 0 | loss 4.3974 | lr 1.30e-03 | grad 0.3457
+2026-04-08 14:03:07 - INFO - parrotllm.training - step    590 | epoch 0 | loss 4.5168 | lr 1.30e-03 | grad 0.5359
+2026-04-08 14:03:10 - INFO - parrotllm.training - step    600 | epoch 0 | loss 4.5400 | lr 1.30e-03 | grad 0.3343
+2026-04-08 14:03:10 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:03:10 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:03:16 - INFO - parrotllm.training -   Train: loss=4.5400, ppl=93.69
+2026-04-08 14:03:16 - INFO - parrotllm.training -   Val:   loss=4.5068, ppl=90.63
+2026-04-08 14:03:16 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:03:17 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p5068_epoch_0000_step_0000600.pt
+2026-04-08 14:03:17 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:03:20 - INFO - parrotllm.training - step    610 | epoch 0 | loss 4.4753 | lr 1.30e-03 | grad 0.2855
+2026-04-08 14:03:24 - INFO - parrotllm.training - step    620 | epoch 0 | loss 4.4947 | lr 1.30e-03 | grad 0.3025
+2026-04-08 14:03:27 - INFO - parrotllm.training - step    630 | epoch 0 | loss 4.4394 | lr 1.30e-03 | grad 0.4398
+2026-04-08 14:03:30 - INFO - parrotllm.training - step    640 | epoch 0 | loss 4.4087 | lr 1.30e-03 | grad 0.2784
+2026-04-08 14:03:33 - INFO - parrotllm.training - step    650 | epoch 0 | loss 4.4705 | lr 1.30e-03 | grad 0.2882
+2026-04-08 14:03:36 - INFO - parrotllm.training - step    660 | epoch 0 | loss 4.3001 | lr 1.30e-03 | grad 0.4116
+2026-04-08 14:03:39 - INFO - parrotllm.training - step    670 | epoch 0 | loss 4.4093 | lr 1.30e-03 | grad 0.3227
+2026-04-08 14:03:42 - INFO - parrotllm.training - step    680 | epoch 0 | loss 4.4190 | lr 1.30e-03 | grad 0.3476
+2026-04-08 14:03:45 - INFO - parrotllm.training - step    690 | epoch 0 | loss 4.2442 | lr 1.30e-03 | grad 0.3783
+2026-04-08 14:03:48 - INFO - parrotllm.training - step    700 | epoch 0 | loss 4.3616 | lr 1.30e-03 | grad 0.5368
+2026-04-08 14:03:48 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:03:48 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:03:54 - INFO - parrotllm.training -   Train: loss=4.3616, ppl=78.39
+2026-04-08 14:03:54 - INFO - parrotllm.training -   Val:   loss=4.3528, ppl=77.70
+2026-04-08 14:03:54 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:03:55 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p3528_epoch_0000_step_0000700.pt
+2026-04-08 14:03:56 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:03:59 - INFO - parrotllm.training - step    710 | epoch 0 | loss 4.2724 | lr 1.30e-03 | grad 0.2602
+2026-04-08 14:04:02 - INFO - parrotllm.training - step    720 | epoch 0 | loss 4.3498 | lr 1.30e-03 | grad 0.3388
+2026-04-08 14:04:05 - INFO - parrotllm.training - step    730 | epoch 0 | loss 4.2156 | lr 1.30e-03 | grad 0.3329
+2026-04-08 14:04:08 - INFO - parrotllm.training - step    740 | epoch 0 | loss 4.2054 | lr 1.30e-03 | grad 0.2879
+2026-04-08 14:04:11 - INFO - parrotllm.training - step    750 | epoch 0 | loss 4.2538 | lr 1.30e-03 | grad 0.3625
+2026-04-08 14:04:12 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0000750.pt
+2026-04-08 14:04:16 - INFO - parrotllm.training - step    760 | epoch 0 | loss 4.2472 | lr 1.30e-03 | grad 0.2917
+2026-04-08 14:04:19 - INFO - parrotllm.training - step    770 | epoch 0 | loss 4.1726 | lr 1.30e-03 | grad 0.3386
+2026-04-08 14:04:22 - INFO - parrotllm.training - step    780 | epoch 0 | loss 4.2424 | lr 1.30e-03 | grad 0.2958
+2026-04-08 14:04:25 - INFO - parrotllm.training - step    790 | epoch 0 | loss 4.2461 | lr 1.30e-03 | grad 0.3299
+2026-04-08 14:04:28 - INFO - parrotllm.training - step    800 | epoch 0 | loss 4.2986 | lr 1.30e-03 | grad 0.3326
+2026-04-08 14:04:28 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:04:28 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:04:34 - INFO - parrotllm.training -   Train: loss=4.2986, ppl=73.60
+2026-04-08 14:04:34 - INFO - parrotllm.training -   Val:   loss=4.2289, ppl=68.64
+2026-04-08 14:04:34 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:04:35 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p2289_epoch_0000_step_0000800.pt
+2026-04-08 14:04:36 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:04:39 - INFO - parrotllm.training - step    810 | epoch 0 | loss 4.1100 | lr 1.30e-03 | grad 0.3297
+2026-04-08 14:04:42 - INFO - parrotllm.training - step    820 | epoch 0 | loss 4.1684 | lr 1.30e-03 | grad 0.4951
+2026-04-08 14:04:45 - INFO - parrotllm.training - step    830 | epoch 0 | loss 4.2305 | lr 1.30e-03 | grad 0.2999
+2026-04-08 14:04:48 - INFO - parrotllm.training - step    840 | epoch 0 | loss 4.1809 | lr 1.30e-03 | grad 0.2789
+2026-04-08 14:04:51 - INFO - parrotllm.training - step    850 | epoch 0 | loss 4.1167 | lr 1.30e-03 | grad 0.3522
+2026-04-08 14:04:54 - INFO - parrotllm.training - step    860 | epoch 0 | loss 4.0878 | lr 1.30e-03 | grad 0.3210
+2026-04-08 14:04:57 - INFO - parrotllm.training - step    870 | epoch 0 | loss 4.1254 | lr 1.30e-03 | grad 0.3781
+2026-04-08 14:05:01 - INFO - parrotllm.training - step    880 | epoch 0 | loss 4.2074 | lr 1.30e-03 | grad 0.3715
+2026-04-08 14:05:04 - INFO - parrotllm.training - step    890 | epoch 0 | loss 4.1691 | lr 1.30e-03 | grad 0.3186
+2026-04-08 14:05:07 - INFO - parrotllm.training - step    900 | epoch 0 | loss 4.1067 | lr 1.30e-03 | grad 0.3043
+2026-04-08 14:05:07 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:05:07 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:05:13 - INFO - parrotllm.training -   Train: loss=4.1067, ppl=60.75
+2026-04-08 14:05:13 - INFO - parrotllm.training -   Val:   loss=4.1455, ppl=63.15
+2026-04-08 14:05:13 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:05:14 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p1455_epoch_0000_step_0000900.pt
+2026-04-08 14:05:14 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:05:17 - INFO - parrotllm.training - step    910 | epoch 0 | loss 4.1516 | lr 1.30e-03 | grad 0.3303
+2026-04-08 14:05:20 - INFO - parrotllm.training - step    920 | epoch 0 | loss 4.1988 | lr 1.30e-03 | grad 0.2828
+2026-04-08 14:05:23 - INFO - parrotllm.training - step    930 | epoch 0 | loss 4.1395 | lr 1.30e-03 | grad 0.2683
+2026-04-08 14:05:27 - INFO - parrotllm.training - step    940 | epoch 0 | loss 4.1705 | lr 1.30e-03 | grad 0.2802
+2026-04-08 14:05:30 - INFO - parrotllm.training - step    950 | epoch 0 | loss 4.1899 | lr 1.30e-03 | grad 0.2848
+2026-04-08 14:05:33 - INFO - parrotllm.training - step    960 | epoch 0 | loss 4.0292 | lr 1.30e-03 | grad 0.3756
+2026-04-08 14:05:36 - INFO - parrotllm.training - step    970 | epoch 0 | loss 4.2175 | lr 1.30e-03 | grad 0.3247
+2026-04-08 14:05:39 - INFO - parrotllm.training - step    980 | epoch 0 | loss 4.0566 | lr 1.30e-03 | grad 0.2580
+2026-04-08 14:05:42 - INFO - parrotllm.training - step    990 | epoch 0 | loss 3.9257 | lr 1.30e-03 | grad 0.3235
+2026-04-08 14:05:45 - INFO - parrotllm.training - step   1000 | epoch 0 | loss 4.1197 | lr 1.30e-03 | grad 0.3498
+2026-04-08 14:05:45 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:05:45 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:05:51 - INFO - parrotllm.training -   Train: loss=4.1197, ppl=61.54
+2026-04-08 14:05:51 - INFO - parrotllm.training -   Val:   loss=4.0852, ppl=59.46
+2026-04-08 14:05:51 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:05:52 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p0852_epoch_0000_step_0001000.pt
+2026-04-08 14:05:52 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:05:54 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0001000.pt
+2026-04-08 14:05:57 - INFO - parrotllm.training - step   1010 | epoch 0 | loss 3.9842 | lr 1.30e-03 | grad 0.2686
+2026-04-08 14:06:01 - INFO - parrotllm.training - step   1020 | epoch 0 | loss 4.0029 | lr 1.30e-03 | grad 0.3144
+2026-04-08 14:06:04 - INFO - parrotllm.training - step   1030 | epoch 0 | loss 4.0865 | lr 1.30e-03 | grad 0.3201
+2026-04-08 14:06:07 - INFO - parrotllm.training - step   1040 | epoch 0 | loss 4.0908 | lr 1.30e-03 | grad 0.2985
+2026-04-08 14:06:10 - INFO - parrotllm.training - step   1050 | epoch 0 | loss 4.0157 | lr 1.30e-03 | grad 0.2554
+2026-04-08 14:06:13 - INFO - parrotllm.training - step   1060 | epoch 0 | loss 4.0633 | lr 1.30e-03 | grad 0.2987
+2026-04-08 14:06:16 - INFO - parrotllm.training - step   1070 | epoch 0 | loss 4.1145 | lr 1.30e-03 | grad 0.3536
+2026-04-08 14:06:19 - INFO - parrotllm.training - step   1080 | epoch 0 | loss 4.1752 | lr 1.30e-03 | grad 0.3480
+2026-04-08 14:06:22 - INFO - parrotllm.training - step   1090 | epoch 0 | loss 4.0568 | lr 1.30e-03 | grad 0.2606
+2026-04-08 14:06:25 - INFO - parrotllm.training - step   1100 | epoch 0 | loss 3.9836 | lr 1.30e-03 | grad 0.2893
+2026-04-08 14:06:25 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:06:25 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:06:31 - INFO - parrotllm.training -   Train: loss=3.9836, ppl=53.71
+2026-04-08 14:06:31 - INFO - parrotllm.training -   Val:   loss=4.0299, ppl=56.25
+2026-04-08 14:06:31 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:06:32 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_4p0299_epoch_0000_step_0001100.pt
+2026-04-08 14:06:33 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:06:36 - INFO - parrotllm.training - step   1110 | epoch 0 | loss 3.9808 | lr 1.30e-03 | grad 0.3244
+2026-04-08 14:06:39 - INFO - parrotllm.training - step   1120 | epoch 0 | loss 3.9793 | lr 1.30e-03 | grad 0.3159
+2026-04-08 14:06:42 - INFO - parrotllm.training - step   1130 | epoch 0 | loss 4.0339 | lr 1.30e-03 | grad 0.3257
+2026-04-08 14:06:45 - INFO - parrotllm.training - step   1140 | epoch 0 | loss 4.0003 | lr 1.30e-03 | grad 0.3093
+2026-04-08 14:06:48 - INFO - parrotllm.training - step   1150 | epoch 0 | loss 4.0389 | lr 1.30e-03 | grad 0.2516
+2026-04-08 14:06:51 - INFO - parrotllm.training - step   1160 | epoch 0 | loss 3.9720 | lr 1.30e-03 | grad 0.2786
+2026-04-08 14:06:54 - INFO - parrotllm.training - step   1170 | epoch 0 | loss 4.0468 | lr 1.30e-03 | grad 0.3113
+2026-04-08 14:06:57 - INFO - parrotllm.training - step   1180 | epoch 0 | loss 3.9364 | lr 1.30e-03 | grad 0.2805
+2026-04-08 14:07:00 - INFO - parrotllm.training - step   1190 | epoch 0 | loss 3.9555 | lr 1.30e-03 | grad 0.2644
+2026-04-08 14:07:03 - INFO - parrotllm.training - step   1200 | epoch 0 | loss 4.0154 | lr 1.30e-03 | grad 0.2491
+2026-04-08 14:07:03 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:07:03 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:07:09 - INFO - parrotllm.training -   Train: loss=4.0154, ppl=55.45
+2026-04-08 14:07:09 - INFO - parrotllm.training -   Val:   loss=3.9848, ppl=53.78
+2026-04-08 14:07:09 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:07:10 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p9848_epoch_0000_step_0001200.pt
+2026-04-08 14:07:11 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:07:14 - INFO - parrotllm.training - step   1210 | epoch 0 | loss 3.9700 | lr 1.30e-03 | grad 0.3052
+2026-04-08 14:07:17 - INFO - parrotllm.training - step   1220 | epoch 0 | loss 3.9788 | lr 1.30e-03 | grad 0.2575
+2026-04-08 14:07:20 - INFO - parrotllm.training - step   1230 | epoch 0 | loss 3.9392 | lr 1.30e-03 | grad 0.2760
+2026-04-08 14:07:23 - INFO - parrotllm.training - step   1240 | epoch 0 | loss 3.8777 | lr 1.30e-03 | grad 0.2860
+2026-04-08 14:07:26 - INFO - parrotllm.training - step   1250 | epoch 0 | loss 3.9115 | lr 1.30e-03 | grad 0.2971
+2026-04-08 14:07:28 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0001250.pt
+2026-04-08 14:07:31 - INFO - parrotllm.training - step   1260 | epoch 0 | loss 3.7619 | lr 1.30e-03 | grad 0.2587
+2026-04-08 14:07:35 - INFO - parrotllm.training - step   1270 | epoch 0 | loss 3.9177 | lr 1.30e-03 | grad 0.3031
+2026-04-08 14:07:38 - INFO - parrotllm.training - step   1280 | epoch 0 | loss 3.8730 | lr 1.30e-03 | grad 0.2635
+2026-04-08 14:07:41 - INFO - parrotllm.training - step   1290 | epoch 0 | loss 3.9729 | lr 1.30e-03 | grad 0.2894
+2026-04-08 14:07:44 - INFO - parrotllm.training - step   1300 | epoch 0 | loss 3.8600 | lr 1.30e-03 | grad 0.2936
+2026-04-08 14:07:44 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:07:44 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:07:50 - INFO - parrotllm.training -   Train: loss=3.8600, ppl=47.46
+2026-04-08 14:07:50 - INFO - parrotllm.training -   Val:   loss=3.9564, ppl=52.27
+2026-04-08 14:07:50 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:07:51 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p9564_epoch_0000_step_0001300.pt
+2026-04-08 14:07:52 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:07:55 - INFO - parrotllm.training - step   1310 | epoch 0 | loss 3.8823 | lr 1.30e-03 | grad 0.2928
+2026-04-08 14:07:58 - INFO - parrotllm.training - step   1320 | epoch 0 | loss 3.9341 | lr 1.30e-03 | grad 0.2853
+2026-04-08 14:08:01 - INFO - parrotllm.training - step   1330 | epoch 0 | loss 3.8283 | lr 1.30e-03 | grad 0.2420
+2026-04-08 14:08:04 - INFO - parrotllm.training - step   1340 | epoch 0 | loss 3.9674 | lr 1.30e-03 | grad 0.2990
+2026-04-08 14:08:07 - INFO - parrotllm.training - step   1350 | epoch 0 | loss 3.8414 | lr 1.30e-03 | grad 0.2488
+2026-04-08 14:08:10 - INFO - parrotllm.training - step   1360 | epoch 0 | loss 3.9046 | lr 1.30e-03 | grad 0.2937
+2026-04-08 14:08:13 - INFO - parrotllm.training - step   1370 | epoch 0 | loss 3.9838 | lr 1.30e-03 | grad 0.2599
+2026-04-08 14:08:16 - INFO - parrotllm.training - step   1380 | epoch 0 | loss 3.7984 | lr 1.30e-03 | grad 0.2948
+2026-04-08 14:08:19 - INFO - parrotllm.training - step   1390 | epoch 0 | loss 4.0469 | lr 1.30e-03 | grad 0.3540
+2026-04-08 14:08:22 - INFO - parrotllm.training - step   1400 | epoch 0 | loss 3.7778 | lr 1.30e-03 | grad 0.3137
+2026-04-08 14:08:22 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:08:22 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:08:28 - INFO - parrotllm.training -   Train: loss=3.7778, ppl=43.72
+2026-04-08 14:08:28 - INFO - parrotllm.training -   Val:   loss=3.9309, ppl=50.95
+2026-04-08 14:08:28 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:08:29 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p9309_epoch_0000_step_0001400.pt
+2026-04-08 14:08:30 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:08:33 - INFO - parrotllm.training - step   1410 | epoch 0 | loss 3.7954 | lr 1.30e-03 | grad 0.2956
+2026-04-08 14:08:36 - INFO - parrotllm.training - step   1420 | epoch 0 | loss 3.9411 | lr 1.30e-03 | grad 0.2454
+2026-04-08 14:08:39 - INFO - parrotllm.training - step   1430 | epoch 0 | loss 3.8850 | lr 1.30e-03 | grad 0.2501
+2026-04-08 14:08:42 - INFO - parrotllm.training - step   1440 | epoch 0 | loss 3.9137 | lr 1.30e-03 | grad 0.2876
+2026-04-08 14:08:45 - INFO - parrotllm.training - step   1450 | epoch 0 | loss 3.8554 | lr 1.30e-03 | grad 0.3139
+2026-04-08 14:08:49 - INFO - parrotllm.training - step   1460 | epoch 0 | loss 3.9747 | lr 1.30e-03 | grad 0.2544
+2026-04-08 14:08:52 - INFO - parrotllm.training - step   1470 | epoch 0 | loss 3.9130 | lr 1.30e-03 | grad 0.2749
+2026-04-08 14:08:55 - INFO - parrotllm.training - step   1480 | epoch 0 | loss 3.8467 | lr 1.30e-03 | grad 0.2734
+2026-04-08 14:08:58 - INFO - parrotllm.training - step   1490 | epoch 0 | loss 3.8300 | lr 1.30e-03 | grad 0.2944
+2026-04-08 14:09:01 - INFO - parrotllm.training - step   1500 | epoch 0 | loss 3.8254 | lr 1.30e-03 | grad 0.2527
+2026-04-08 14:09:01 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:09:01 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:09:07 - INFO - parrotllm.training -   Train: loss=3.8254, ppl=45.85
+2026-04-08 14:09:07 - INFO - parrotllm.training -   Val:   loss=3.8977, ppl=49.29
+2026-04-08 14:09:07 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:09:08 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p8977_epoch_0000_step_0001500.pt
+2026-04-08 14:09:08 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:09:09 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0001500.pt
+2026-04-08 14:09:13 - INFO - parrotllm.training - step   1510 | epoch 0 | loss 3.9278 | lr 1.30e-03 | grad 0.2797
+2026-04-08 14:09:17 - INFO - parrotllm.training - step   1520 | epoch 0 | loss 3.7788 | lr 1.30e-03 | grad 0.2787
+2026-04-08 14:09:20 - INFO - parrotllm.training - step   1530 | epoch 0 | loss 3.8734 | lr 1.30e-03 | grad 0.2912
+2026-04-08 14:09:23 - INFO - parrotllm.training - step   1540 | epoch 0 | loss 3.8463 | lr 1.30e-03 | grad 0.2847
+2026-04-08 14:09:26 - INFO - parrotllm.training - step   1550 | epoch 0 | loss 3.8634 | lr 1.30e-03 | grad 0.2400
+2026-04-08 14:09:29 - INFO - parrotllm.training - step   1560 | epoch 0 | loss 3.8797 | lr 1.30e-03 | grad 0.2845
+2026-04-08 14:09:32 - INFO - parrotllm.training - step   1570 | epoch 0 | loss 3.8963 | lr 1.30e-03 | grad 0.3677
+2026-04-08 14:09:35 - INFO - parrotllm.training - step   1580 | epoch 0 | loss 3.8304 | lr 1.30e-03 | grad 0.2441
+2026-04-08 14:09:38 - INFO - parrotllm.training - step   1590 | epoch 0 | loss 3.8104 | lr 1.30e-03 | grad 0.2965
+2026-04-08 14:09:41 - INFO - parrotllm.training - step   1600 | epoch 0 | loss 3.9313 | lr 1.30e-03 | grad 0.2674
+2026-04-08 14:09:41 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:09:41 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:09:47 - INFO - parrotllm.training -   Train: loss=3.9313, ppl=50.97
+2026-04-08 14:09:47 - INFO - parrotllm.training -   Val:   loss=3.8751, ppl=48.19
+2026-04-08 14:09:47 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:09:48 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p8751_epoch_0000_step_0001600.pt
+2026-04-08 14:09:49 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:09:52 - INFO - parrotllm.training - step   1610 | epoch 0 | loss 3.7593 | lr 1.30e-03 | grad 0.2693
+2026-04-08 14:09:55 - INFO - parrotllm.training - step   1620 | epoch 0 | loss 3.7896 | lr 1.30e-03 | grad 0.2712
+2026-04-08 14:09:58 - INFO - parrotllm.training - step   1630 | epoch 0 | loss 3.8531 | lr 1.30e-03 | grad 0.3044
+2026-04-08 14:10:01 - INFO - parrotllm.training - step   1640 | epoch 0 | loss 3.8099 | lr 1.30e-03 | grad 0.2867
+2026-04-08 14:10:04 - INFO - parrotllm.training - step   1650 | epoch 0 | loss 3.7450 | lr 1.30e-03 | grad 0.2358
+2026-04-08 14:10:07 - INFO - parrotllm.training - step   1660 | epoch 0 | loss 3.8514 | lr 1.30e-03 | grad 0.2959
+2026-04-08 14:10:10 - INFO - parrotllm.training - step   1670 | epoch 0 | loss 3.9372 | lr 1.30e-03 | grad 0.2563
+2026-04-08 14:10:13 - INFO - parrotllm.training - step   1680 | epoch 0 | loss 3.8601 | lr 1.30e-03 | grad 0.2509
+2026-04-08 14:10:16 - INFO - parrotllm.training - step   1690 | epoch 0 | loss 3.8107 | lr 1.30e-03 | grad 0.3501
+2026-04-08 14:10:19 - INFO - parrotllm.training - step   1700 | epoch 0 | loss 3.8500 | lr 1.30e-03 | grad 0.2807
+2026-04-08 14:10:19 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:10:19 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:10:25 - INFO - parrotllm.training -   Train: loss=3.8500, ppl=46.99
+2026-04-08 14:10:25 - INFO - parrotllm.training -   Val:   loss=3.8609, ppl=47.51
+2026-04-08 14:10:25 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:10:26 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p8609_epoch_0000_step_0001700.pt
+2026-04-08 14:10:27 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:10:30 - INFO - parrotllm.training - step   1710 | epoch 0 | loss 3.8379 | lr 1.30e-03 | grad 0.2640
+2026-04-08 14:10:33 - INFO - parrotllm.training - step   1720 | epoch 0 | loss 3.7781 | lr 1.30e-03 | grad 0.2220
+2026-04-08 14:10:36 - INFO - parrotllm.training - step   1730 | epoch 0 | loss 3.7862 | lr 1.30e-03 | grad 0.2274
+2026-04-08 14:10:39 - INFO - parrotllm.training - step   1740 | epoch 0 | loss 3.7407 | lr 1.30e-03 | grad 0.2540
+2026-04-08 14:10:42 - INFO - parrotllm.training - step   1750 | epoch 0 | loss 3.7586 | lr 1.30e-03 | grad 0.2440
+2026-04-08 14:10:43 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0000_step_0001750.pt
+2026-04-08 14:10:47 - INFO - parrotllm.training - step   1760 | epoch 0 | loss 3.9668 | lr 1.30e-03 | grad 0.2548
+2026-04-08 14:10:51 - INFO - parrotllm.training - step   1770 | epoch 0 | loss 3.8277 | lr 1.30e-03 | grad 0.3181
+2026-04-08 14:10:54 - INFO - parrotllm.training - step   1780 | epoch 0 | loss 3.7309 | lr 1.30e-03 | grad 0.2717
+2026-04-08 14:10:57 - INFO - parrotllm.training - step   1790 | epoch 0 | loss 3.8077 | lr 1.30e-03 | grad 0.2392
+2026-04-08 14:11:00 - INFO - parrotllm.training - step   1800 | epoch 0 | loss 3.7557 | lr 1.30e-03 | grad 0.2802
+2026-04-08 14:11:00 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:11:00 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:11:06 - INFO - parrotllm.training -   Train: loss=3.7557, ppl=42.76
+2026-04-08 14:11:06 - INFO - parrotllm.training -   Val:   loss=3.8397, ppl=46.51
+2026-04-08 14:11:06 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:11:07 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p8397_epoch_0000_step_0001800.pt
+2026-04-08 14:11:08 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:11:11 - INFO - parrotllm.training - step   1810 | epoch 0 | loss 3.8761 | lr 1.30e-03 | grad 0.2626
+2026-04-08 14:11:14 - INFO - parrotllm.training - step   1820 | epoch 0 | loss 3.7974 | lr 1.30e-03 | grad 0.2726
+2026-04-08 14:11:17 - INFO - parrotllm.training - step   1830 | epoch 0 | loss 3.8158 | lr 1.30e-03 | grad 0.2500
+2026-04-08 14:11:20 - INFO - parrotllm.training - step   1840 | epoch 0 | loss 3.7507 | lr 1.30e-03 | grad 0.3041
+2026-04-08 14:11:23 - INFO - parrotllm.training - step   1850 | epoch 0 | loss 3.7376 | lr 1.30e-03 | grad 0.2492
+2026-04-08 14:11:26 - INFO - parrotllm.training - step   1860 | epoch 0 | loss 3.7277 | lr 1.30e-03 | grad 0.3965
+2026-04-08 14:11:29 - INFO - parrotllm.training - step   1870 | epoch 0 | loss 3.9267 | lr 1.30e-03 | grad 0.2600
+2026-04-08 14:11:32 - INFO - parrotllm.training - step   1880 | epoch 0 | loss 3.8510 | lr 1.30e-03 | grad 0.2390
+2026-04-08 14:11:35 - INFO - parrotllm.training - step   1890 | epoch 0 | loss 3.8086 | lr 1.30e-03 | grad 0.2429
+2026-04-08 14:11:38 - INFO - parrotllm.training - step   1900 | epoch 0 | loss 3.7977 | lr 1.30e-03 | grad 0.2865
+2026-04-08 14:11:38 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:11:38 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:11:44 - INFO - parrotllm.training -   Train: loss=3.7977, ppl=44.60
+2026-04-08 14:11:44 - INFO - parrotllm.training -   Val:   loss=3.8203, ppl=45.62
+2026-04-08 14:11:44 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:11:45 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p8203_epoch_0000_step_0001900.pt
+2026-04-08 14:11:46 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:11:49 - INFO - parrotllm.training - step   1910 | epoch 0 | loss 3.7643 | lr 1.30e-03 | grad 0.2595
+2026-04-08 14:11:52 - INFO - parrotllm.training - step   1920 | epoch 0 | loss 3.8758 | lr 1.30e-03 | grad 0.2901
+2026-04-08 14:11:55 - INFO - parrotllm.training - step   1930 | epoch 0 | loss 3.8156 | lr 1.30e-03 | grad 0.2178
+2026-04-08 14:11:58 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:11:58 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:12:04 - INFO - parrotllm.training - Epoch 1 complete:
+2026-04-08 14:12:04 - INFO - parrotllm.training -   Train: loss=3.7150, ppl=41.06
+2026-04-08 14:12:04 - INFO - parrotllm.training -   Val:   loss=3.8175, ppl=45.49
+2026-04-08 14:12:04 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:12:05 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p8175_epoch_0001_step_0001939.pt
+2026-04-08 14:12:06 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:12:06 - INFO - parrotllm.training - step   1940 | epoch 1 | loss 3.7520 | lr 1.30e-03 | grad 0.2686
+2026-04-08 14:12:09 - INFO - parrotllm.training - step   1950 | epoch 1 | loss 3.6451 | lr 1.30e-03 | grad 0.3257
+2026-04-08 14:12:12 - INFO - parrotllm.training - step   1960 | epoch 1 | loss 3.8351 | lr 1.30e-03 | grad 0.2382
+2026-04-08 14:12:15 - INFO - parrotllm.training - step   1970 | epoch 1 | loss 3.6507 | lr 1.30e-03 | grad 0.2813
+2026-04-08 14:12:18 - INFO - parrotllm.training - step   1980 | epoch 1 | loss 3.7334 | lr 1.30e-03 | grad 0.2410
+2026-04-08 14:12:22 - INFO - parrotllm.training - step   1990 | epoch 1 | loss 3.8006 | lr 1.30e-03 | grad 0.2702
+2026-04-08 14:12:25 - INFO - parrotllm.training - step   2000 | epoch 1 | loss 3.6891 | lr 1.30e-03 | grad 0.2686
+2026-04-08 14:12:25 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:12:25 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:12:31 - INFO - parrotllm.training -   Train: loss=3.6891, ppl=40.01
+2026-04-08 14:12:31 - INFO - parrotllm.training -   Val:   loss=3.8102, ppl=45.16
+2026-04-08 14:12:31 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:12:32 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p8102_epoch_0001_step_0002000.pt
+2026-04-08 14:12:32 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:12:33 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0002000.pt
+2026-04-08 14:12:37 - INFO - parrotllm.training - step   2010 | epoch 1 | loss 3.8102 | lr 1.30e-03 | grad 0.2732
+2026-04-08 14:12:40 - INFO - parrotllm.training - step   2020 | epoch 1 | loss 3.7550 | lr 1.30e-03 | grad 0.2384
+2026-04-08 14:12:43 - INFO - parrotllm.training - step   2030 | epoch 1 | loss 3.8270 | lr 1.30e-03 | grad 0.2987
+2026-04-08 14:12:46 - INFO - parrotllm.training - step   2040 | epoch 1 | loss 3.7666 | lr 1.30e-03 | grad 0.2333
+2026-04-08 14:12:50 - INFO - parrotllm.training - step   2050 | epoch 1 | loss 3.6323 | lr 1.30e-03 | grad 0.2599
+2026-04-08 14:12:53 - INFO - parrotllm.training - step   2060 | epoch 1 | loss 3.6997 | lr 1.30e-03 | grad 0.2384
+2026-04-08 14:12:56 - INFO - parrotllm.training - step   2070 | epoch 1 | loss 3.6985 | lr 1.30e-03 | grad 0.3011
+2026-04-08 14:12:59 - INFO - parrotllm.training - step   2080 | epoch 1 | loss 3.7032 | lr 1.30e-03 | grad 0.2436
+2026-04-08 14:13:02 - INFO - parrotllm.training - step   2090 | epoch 1 | loss 3.8128 | lr 1.30e-03 | grad 0.2806
+2026-04-08 14:13:05 - INFO - parrotllm.training - step   2100 | epoch 1 | loss 3.7440 | lr 1.30e-03 | grad 0.2722
+2026-04-08 14:13:05 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:13:05 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:13:11 - INFO - parrotllm.training -   Train: loss=3.7440, ppl=42.26
+2026-04-08 14:13:11 - INFO - parrotllm.training -   Val:   loss=3.7972, ppl=44.58
+2026-04-08 14:13:11 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:13:12 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7972_epoch_0001_step_0002100.pt
+2026-04-08 14:13:12 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:13:15 - INFO - parrotllm.training - step   2110 | epoch 1 | loss 3.6330 | lr 1.30e-03 | grad 0.2388
+2026-04-08 14:13:18 - INFO - parrotllm.training - step   2120 | epoch 1 | loss 3.7768 | lr 1.30e-03 | grad 0.2246
+2026-04-08 14:13:22 - INFO - parrotllm.training - step   2130 | epoch 1 | loss 3.8282 | lr 1.30e-03 | grad 0.2792
+2026-04-08 14:13:25 - INFO - parrotllm.training - step   2140 | epoch 1 | loss 3.8310 | lr 1.30e-03 | grad 0.2753
+2026-04-08 14:13:28 - INFO - parrotllm.training - step   2150 | epoch 1 | loss 3.6568 | lr 1.30e-03 | grad 0.2549
+2026-04-08 14:13:31 - INFO - parrotllm.training - step   2160 | epoch 1 | loss 3.7285 | lr 1.30e-03 | grad 0.3225
+2026-04-08 14:13:34 - INFO - parrotllm.training - step   2170 | epoch 1 | loss 3.7505 | lr 1.30e-03 | grad 0.2690
+2026-04-08 14:13:37 - INFO - parrotllm.training - step   2180 | epoch 1 | loss 3.7960 | lr 1.30e-03 | grad 0.2279
+2026-04-08 14:13:40 - INFO - parrotllm.training - step   2190 | epoch 1 | loss 3.7219 | lr 1.30e-03 | grad 0.2701
+2026-04-08 14:13:43 - INFO - parrotllm.training - step   2200 | epoch 1 | loss 3.7313 | lr 1.30e-03 | grad 0.2755
+2026-04-08 14:13:43 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:13:43 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:13:49 - INFO - parrotllm.training -   Train: loss=3.7313, ppl=41.74
+2026-04-08 14:13:49 - INFO - parrotllm.training -   Val:   loss=3.7865, ppl=44.10
+2026-04-08 14:13:49 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:13:50 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7865_epoch_0001_step_0002200.pt
+2026-04-08 14:13:50 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:13:53 - INFO - parrotllm.training - step   2210 | epoch 1 | loss 3.7195 | lr 1.30e-03 | grad 0.2528
+2026-04-08 14:13:57 - INFO - parrotllm.training - step   2220 | epoch 1 | loss 3.7649 | lr 1.30e-03 | grad 0.3107
+2026-04-08 14:14:00 - INFO - parrotllm.training - step   2230 | epoch 1 | loss 3.7344 | lr 1.30e-03 | grad 0.2261
+2026-04-08 14:14:03 - INFO - parrotllm.training - step   2240 | epoch 1 | loss 3.7536 | lr 1.30e-03 | grad 0.2431
+2026-04-08 14:14:06 - INFO - parrotllm.training - step   2250 | epoch 1 | loss 3.8898 | lr 1.30e-03 | grad 0.2686
+2026-04-08 14:14:07 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0002250.pt
+2026-04-08 14:14:11 - INFO - parrotllm.training - step   2260 | epoch 1 | loss 3.7875 | lr 1.30e-03 | grad 0.2420
+2026-04-08 14:14:14 - INFO - parrotllm.training - step   2270 | epoch 1 | loss 3.6924 | lr 1.30e-03 | grad 0.3212
+2026-04-08 14:14:17 - INFO - parrotllm.training - step   2280 | epoch 1 | loss 3.6053 | lr 1.30e-03 | grad 0.2359
+2026-04-08 14:14:20 - INFO - parrotllm.training - step   2290 | epoch 1 | loss 3.7718 | lr 1.30e-03 | grad 0.2407
+2026-04-08 14:14:23 - INFO - parrotllm.training - step   2300 | epoch 1 | loss 3.6741 | lr 1.30e-03 | grad 0.2590
+2026-04-08 14:14:23 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:14:23 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:14:29 - INFO - parrotllm.training -   Train: loss=3.6741, ppl=39.41
+2026-04-08 14:14:29 - INFO - parrotllm.training -   Val:   loss=3.7749, ppl=43.59
+2026-04-08 14:14:29 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:14:30 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7749_epoch_0001_step_0002300.pt
+2026-04-08 14:14:31 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:14:34 - INFO - parrotllm.training - step   2310 | epoch 1 | loss 3.6128 | lr 1.30e-03 | grad 0.2328
+2026-04-08 14:14:37 - INFO - parrotllm.training - step   2320 | epoch 1 | loss 3.7302 | lr 1.30e-03 | grad 0.2520
+2026-04-08 14:14:40 - INFO - parrotllm.training - step   2330 | epoch 1 | loss 3.7759 | lr 1.30e-03 | grad 0.3411
+2026-04-08 14:14:43 - INFO - parrotllm.training - step   2340 | epoch 1 | loss 3.7055 | lr 1.30e-03 | grad 0.2889
+2026-04-08 14:14:46 - INFO - parrotllm.training - step   2350 | epoch 1 | loss 3.7354 | lr 1.30e-03 | grad 0.2489
+2026-04-08 14:14:49 - INFO - parrotllm.training - step   2360 | epoch 1 | loss 3.6530 | lr 1.30e-03 | grad 0.2567
+2026-04-08 14:14:52 - INFO - parrotllm.training - step   2370 | epoch 1 | loss 3.6921 | lr 1.30e-03 | grad 0.3491
+2026-04-08 14:14:55 - INFO - parrotllm.training - step   2380 | epoch 1 | loss 3.6108 | lr 1.30e-03 | grad 0.2613
+2026-04-08 14:14:58 - INFO - parrotllm.training - step   2390 | epoch 1 | loss 3.8232 | lr 1.30e-03 | grad 0.2552
+2026-04-08 14:15:01 - INFO - parrotllm.training - step   2400 | epoch 1 | loss 3.7259 | lr 1.30e-03 | grad 0.2108
+2026-04-08 14:15:01 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:15:01 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:15:07 - INFO - parrotllm.training -   Train: loss=3.7259, ppl=41.51
+2026-04-08 14:15:07 - INFO - parrotllm.training -   Val:   loss=3.7614, ppl=43.01
+2026-04-08 14:15:07 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:15:08 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7614_epoch_0001_step_0002400.pt
+2026-04-08 14:15:09 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:15:12 - INFO - parrotllm.training - step   2410 | epoch 1 | loss 3.7518 | lr 1.30e-03 | grad 0.2214
+2026-04-08 14:15:15 - INFO - parrotllm.training - step   2420 | epoch 1 | loss 3.6875 | lr 1.30e-03 | grad 0.2749
+2026-04-08 14:15:18 - INFO - parrotllm.training - step   2430 | epoch 1 | loss 3.7915 | lr 1.30e-03 | grad 0.2397
+2026-04-08 14:15:21 - INFO - parrotllm.training - step   2440 | epoch 1 | loss 3.7126 | lr 1.30e-03 | grad 0.2261
+2026-04-08 14:15:24 - INFO - parrotllm.training - step   2450 | epoch 1 | loss 3.6834 | lr 1.30e-03 | grad 0.2479
+2026-04-08 14:15:27 - INFO - parrotllm.training - step   2460 | epoch 1 | loss 3.7637 | lr 1.30e-03 | grad 0.2503
+2026-04-08 14:15:30 - INFO - parrotllm.training - step   2470 | epoch 1 | loss 3.7494 | lr 1.30e-03 | grad 0.2728
+2026-04-08 14:15:33 - INFO - parrotllm.training - step   2480 | epoch 1 | loss 3.6249 | lr 1.30e-03 | grad 0.3091
+2026-04-08 14:15:36 - INFO - parrotllm.training - step   2490 | epoch 1 | loss 3.7416 | lr 1.30e-03 | grad 0.2311
+2026-04-08 14:15:39 - INFO - parrotllm.training - step   2500 | epoch 1 | loss 3.7252 | lr 1.30e-03 | grad 0.2412
+2026-04-08 14:15:39 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:15:39 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:15:45 - INFO - parrotllm.training -   Train: loss=3.7252, ppl=41.48
+2026-04-08 14:15:45 - INFO - parrotllm.training -   Val:   loss=3.7530, ppl=42.65
+2026-04-08 14:15:45 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:15:46 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7530_epoch_0001_step_0002500.pt
+2026-04-08 14:15:47 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:15:48 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0002500.pt
+2026-04-08 14:15:52 - INFO - parrotllm.training - step   2510 | epoch 1 | loss 3.8002 | lr 1.30e-03 | grad 0.2592
+2026-04-08 14:15:55 - INFO - parrotllm.training - step   2520 | epoch 1 | loss 3.7339 | lr 1.27e-03 | grad 0.2968
+2026-04-08 14:15:58 - INFO - parrotllm.training - step   2530 | epoch 1 | loss 3.7640 | lr 1.25e-03 | grad 0.2381
+2026-04-08 14:16:01 - INFO - parrotllm.training - step   2540 | epoch 1 | loss 3.5722 | lr 1.23e-03 | grad 0.2414
+2026-04-08 14:16:04 - INFO - parrotllm.training - step   2550 | epoch 1 | loss 3.6808 | lr 1.21e-03 | grad 0.2878
+2026-04-08 14:16:07 - INFO - parrotllm.training - step   2560 | epoch 1 | loss 3.6816 | lr 1.19e-03 | grad 0.2333
+2026-04-08 14:16:10 - INFO - parrotllm.training - step   2570 | epoch 1 | loss 3.7076 | lr 1.17e-03 | grad 0.2260
+2026-04-08 14:16:13 - INFO - parrotllm.training - step   2580 | epoch 1 | loss 3.6863 | lr 1.15e-03 | grad 0.2174
+2026-04-08 14:16:16 - INFO - parrotllm.training - step   2590 | epoch 1 | loss 3.5777 | lr 1.12e-03 | grad 0.2217
+2026-04-08 14:16:19 - INFO - parrotllm.training - step   2600 | epoch 1 | loss 3.7034 | lr 1.10e-03 | grad 0.1942
+2026-04-08 14:16:19 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:16:19 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:16:25 - INFO - parrotllm.training -   Train: loss=3.7034, ppl=40.59
+2026-04-08 14:16:25 - INFO - parrotllm.training -   Val:   loss=3.7307, ppl=41.71
+2026-04-08 14:16:25 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:16:26 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7307_epoch_0001_step_0002600.pt
+2026-04-08 14:16:27 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:16:30 - INFO - parrotllm.training - step   2610 | epoch 1 | loss 3.7031 | lr 1.08e-03 | grad 0.2177
+2026-04-08 14:16:33 - INFO - parrotllm.training - step   2620 | epoch 1 | loss 3.5964 | lr 1.06e-03 | grad 0.2247
+2026-04-08 14:16:36 - INFO - parrotllm.training - step   2630 | epoch 1 | loss 3.6330 | lr 1.04e-03 | grad 0.2196
+2026-04-08 14:16:39 - INFO - parrotllm.training - step   2640 | epoch 1 | loss 3.6384 | lr 1.02e-03 | grad 0.2115
+2026-04-08 14:16:42 - INFO - parrotllm.training - step   2650 | epoch 1 | loss 3.6308 | lr 9.95e-04 | grad 0.2120
+2026-04-08 14:16:45 - INFO - parrotllm.training - step   2660 | epoch 1 | loss 3.7143 | lr 9.74e-04 | grad 0.2094
+2026-04-08 14:16:48 - INFO - parrotllm.training - step   2670 | epoch 1 | loss 3.6325 | lr 9.52e-04 | grad 0.1910
+2026-04-08 14:16:52 - INFO - parrotllm.training - step   2680 | epoch 1 | loss 3.6957 | lr 9.31e-04 | grad 0.2371
+2026-04-08 14:16:55 - INFO - parrotllm.training - step   2690 | epoch 1 | loss 3.5170 | lr 9.09e-04 | grad 0.2324
+2026-04-08 14:16:58 - INFO - parrotllm.training - step   2700 | epoch 1 | loss 3.6375 | lr 8.88e-04 | grad 0.2319
+2026-04-08 14:16:58 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:16:58 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:17:03 - INFO - parrotllm.training -   Train: loss=3.6375, ppl=38.00
+2026-04-08 14:17:03 - INFO - parrotllm.training -   Val:   loss=3.7026, ppl=40.55
+2026-04-08 14:17:03 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:17:04 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p7026_epoch_0001_step_0002700.pt
+2026-04-08 14:17:05 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:17:08 - INFO - parrotllm.training - step   2710 | epoch 1 | loss 3.6604 | lr 8.66e-04 | grad 0.2323
+2026-04-08 14:17:11 - INFO - parrotllm.training - step   2720 | epoch 1 | loss 3.5665 | lr 8.45e-04 | grad 0.2418
+2026-04-08 14:17:14 - INFO - parrotllm.training - step   2730 | epoch 1 | loss 3.6697 | lr 8.23e-04 | grad 0.2085
+2026-04-08 14:17:17 - INFO - parrotllm.training - step   2740 | epoch 1 | loss 3.6119 | lr 8.02e-04 | grad 0.2187
+2026-04-08 14:17:20 - INFO - parrotllm.training - step   2750 | epoch 1 | loss 3.6202 | lr 7.80e-04 | grad 0.2147
+2026-04-08 14:17:21 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0002750.pt
+2026-04-08 14:17:25 - INFO - parrotllm.training - step   2760 | epoch 1 | loss 3.6355 | lr 7.59e-04 | grad 0.2094
+2026-04-08 14:17:28 - INFO - parrotllm.training - step   2770 | epoch 1 | loss 3.5444 | lr 7.37e-04 | grad 0.1881
+2026-04-08 14:17:31 - INFO - parrotllm.training - step   2780 | epoch 1 | loss 3.6314 | lr 7.16e-04 | grad 0.1970
+2026-04-08 14:17:34 - INFO - parrotllm.training - step   2790 | epoch 1 | loss 3.7579 | lr 6.94e-04 | grad 0.1888
+2026-04-08 14:17:38 - INFO - parrotllm.training - step   2800 | epoch 1 | loss 3.5892 | lr 6.73e-04 | grad 0.1976
+2026-04-08 14:17:38 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:17:38 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:17:43 - INFO - parrotllm.training -   Train: loss=3.5892, ppl=36.20
+2026-04-08 14:17:43 - INFO - parrotllm.training -   Val:   loss=3.6796, ppl=39.63
+2026-04-08 14:17:43 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:17:44 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p6796_epoch_0001_step_0002800.pt
+2026-04-08 14:17:45 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:17:48 - INFO - parrotllm.training - step   2810 | epoch 1 | loss 3.5744 | lr 6.51e-04 | grad 0.1830
+2026-04-08 14:17:51 - INFO - parrotllm.training - step   2820 | epoch 1 | loss 3.5782 | lr 6.30e-04 | grad 0.1842
+2026-04-08 14:17:54 - INFO - parrotllm.training - step   2830 | epoch 1 | loss 3.5711 | lr 6.08e-04 | grad 0.1936
+2026-04-08 14:17:58 - INFO - parrotllm.training - step   2840 | epoch 1 | loss 3.6729 | lr 5.87e-04 | grad 0.2174
+2026-04-08 14:18:01 - INFO - parrotllm.training - step   2850 | epoch 1 | loss 3.5268 | lr 5.65e-04 | grad 0.1870
+2026-04-08 14:18:04 - INFO - parrotllm.training - step   2860 | epoch 1 | loss 3.5865 | lr 5.44e-04 | grad 0.1811
+2026-04-08 14:18:07 - INFO - parrotllm.training - step   2870 | epoch 1 | loss 3.8056 | lr 5.22e-04 | grad 0.1992
+2026-04-08 14:18:10 - INFO - parrotllm.training - step   2880 | epoch 1 | loss 3.5511 | lr 5.00e-04 | grad 0.1885
+2026-04-08 14:18:13 - INFO - parrotllm.training - step   2890 | epoch 1 | loss 3.5405 | lr 4.79e-04 | grad 0.1788
+2026-04-08 14:18:16 - INFO - parrotllm.training - step   2900 | epoch 1 | loss 3.5467 | lr 4.57e-04 | grad 0.1684
+2026-04-08 14:18:16 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:18:16 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:18:22 - INFO - parrotllm.training -   Train: loss=3.5467, ppl=34.70
+2026-04-08 14:18:22 - INFO - parrotllm.training -   Val:   loss=3.6558, ppl=38.70
+2026-04-08 14:18:22 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:18:23 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p6558_epoch_0001_step_0002900.pt
+2026-04-08 14:18:23 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:18:27 - INFO - parrotllm.training - step   2910 | epoch 1 | loss 3.4976 | lr 4.36e-04 | grad 0.1702
+2026-04-08 14:18:30 - INFO - parrotllm.training - step   2920 | epoch 1 | loss 3.5874 | lr 4.14e-04 | grad 0.1743
+2026-04-08 14:18:33 - INFO - parrotllm.training - step   2930 | epoch 1 | loss 3.6129 | lr 3.93e-04 | grad 0.1733
+2026-04-08 14:18:36 - INFO - parrotllm.training - step   2940 | epoch 1 | loss 3.6251 | lr 3.71e-04 | grad 0.1962
+2026-04-08 14:18:39 - INFO - parrotllm.training - step   2950 | epoch 1 | loss 3.5845 | lr 3.50e-04 | grad 0.1750
+2026-04-08 14:18:42 - INFO - parrotllm.training - step   2960 | epoch 1 | loss 3.5841 | lr 3.28e-04 | grad 0.1756
+2026-04-08 14:18:45 - INFO - parrotllm.training - step   2970 | epoch 1 | loss 3.5116 | lr 3.07e-04 | grad 0.1809
+2026-04-08 14:18:48 - INFO - parrotllm.training - step   2980 | epoch 1 | loss 3.5941 | lr 2.85e-04 | grad 0.1771
+2026-04-08 14:18:51 - INFO - parrotllm.training - step   2990 | epoch 1 | loss 3.5769 | lr 2.64e-04 | grad 0.1603
+2026-04-08 14:18:54 - INFO - parrotllm.training - step   3000 | epoch 1 | loss 3.5360 | lr 2.42e-04 | grad 0.1627
+2026-04-08 14:18:54 - INFO - parrotllm.training - Starting evaluation...
+2026-04-08 14:18:54 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:19:00 - INFO - parrotllm.training -   Train: loss=3.5360, ppl=34.33
+2026-04-08 14:19:00 - INFO - parrotllm.training -   Val:   loss=3.6338, ppl=37.86
+2026-04-08 14:19:00 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-08 14:19:01 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/best_loss_3p6338_epoch_0001_step_0003000.pt
+2026-04-08 14:19:02 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-08 14:19:03 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0003000.pt
+2026-04-08 14:19:07 - INFO - parrotllm.training - step   3010 | epoch 1 | loss 3.5169 | lr 2.21e-04 | grad 0.1609
+2026-04-08 14:19:10 - INFO - parrotllm.training - step   3020 | epoch 1 | loss 3.5999 | lr 1.99e-04 | grad 0.1654
+2026-04-08 14:19:13 - INFO - parrotllm.training - step   3030 | epoch 1 | loss 3.5634 | lr 1.78e-04 | grad 0.1591
+2026-04-08 14:19:16 - INFO - parrotllm.training - step   3040 | epoch 1 | loss 3.5050 | lr 1.56e-04 | grad 0.1572
+2026-04-08 14:19:19 - INFO - parrotllm.training - step   3050 | epoch 1 | loss 3.5203 | lr 1.35e-04 | grad 0.1517
+2026-04-08 14:19:21 - INFO - parrotllm.training - Saved final checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260408_135803/checkpoints/last_epoch_0001_step_0003052.pt
+2026-04-08 14:19:21 - INFO - parrotllm.training - 
+Loss Curve:
+  Loss
+  ^
+   10.93 |*                                                 
+         |*                                                 
+         |*                                                 
+         |*                                                 
+         |**                                                
+         | ****                                             
+         |    ********                                      
+    3.44 |        ******************************************
+         +--------------------------------------------------> step
+         1                                             3052
+2026-04-08 14:19:21 - INFO - parrotllm.training - 
+============================================================
+TRAINING COMPLETE
+============================================================
+  Epochs: 2
+  Total steps: 3052
+  Total time: 0.35 hours
+  Best validation loss: 3.6338
+  Run directory: runs/big_run/exp_c/run_20260408_135803
+============================================================
+2026-04-08 14:19:21 - INFO - parrotllm.training - Uploading completed run directory to Hugging Face: repo=ParrotLabs/Preprocessed

runs/big_run/exp_c/run_20260408_135803/training_plots.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c3c16bb5b3906d8ba3ff1dd5d122040c63a1b2ab72b16615ef1419a75fd40de1
+size 111349