Upload training run run_20260409_194841

ParrotLLM end-of-training sync for local run runs/big_run/exp_c/run_20260409_194841.

.gitattributes

@@ -62,3 +62,4 @@ runs/big_run/exp_c/run_20260408_124138/training_plots.pdf filter=lfs diff=lfs me
 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_151821/training_plots.pdf filter=lfs diff=lfs merge=lfs -text
 runs/big_run/exp_c/run_20260409_192244/training_plots.pdf filter=lfs diff=lfs merge=lfs -text
+runs/big_run/exp_c/run_20260409_194841/training_plots.pdf filter=lfs diff=lfs merge=lfs -text

runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p4954_epoch_0000_step_0003000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:62df41084669be1f3ee805a393160d23da25ad91ff3e9931ab7ddb1645abeaa3
+size 479869564

runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p5182_epoch_0000_step_0002900.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8dc6825ddee8986b02f27143f3feaec6ce3247c05694afec54d96f35541f8de0
+size 479869564

runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p5390_epoch_0000_step_0002800.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:598960d4f5fc7289d4e99eb50935b17aef06fb01bcc9ba08dbfa27ad3c8ce32e
+size 479869564

runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p5626_epoch_0000_step_0002700.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4f5003313a962076ddd3d31c23361f0afeb6bfccb7a39f2bd722cdf69ad94361
+size 479869564

runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p5883_epoch_0000_step_0002600.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2b92b4da3d8464912056a5749c1bc3ea55535869d452840bb8cc273544729146
+size 479869564

runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p6123_epoch_0000_step_0002500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bee027ccaae42ebf411ed21c5399a6df0c825350ad6d63d9562bbc9a5fcf9971
+size 479869564

runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p6349_epoch_0000_step_0002400.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5f2a7b24fb4cc24d5ee0b6b9d1bd29836a714eadbbc659c2a29c93c1f25175b2
+size 479869564

runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p6574_epoch_0000_step_0002300.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50efdf71d70843697d6d262185fcf050184e43b5d84cdef5577f6a3cad67801e
+size 479869564

runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p6726_epoch_0000_step_0002200.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:84fd8ee7a5367db54383a725dc245e543726359782c559f88d260c4bd8665dae
+size 479869564

runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p6834_epoch_0000_step_0002100.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3758995ece4daeea42e59a092553897792dd90038cad948e507db6b6d380bff2
+size 479869564

runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0001000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e46d5b78ddb72f60a848a08953234725d42ddb051bddd826ddb53ccfa1469f96
+size 479860488

runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0001250.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:82cfab84a61c5e72e054b6df7eb37fd397dd7429672df56064dafb1eac13125c
+size 479860488

runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0001500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4b9b09a194d0cfc3e6712d73f439f70ceab3efa3bb0603fe417a52e35189bb2f
+size 479860488

runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0001750.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9bf624d279361904c48ceab2174b2d194d35e47704714e17b4331b2640d410bb
+size 479860488

runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0002000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60493600f474de78c40a8c382fde3ca79fc14de99ff54569c95024f79d83c288
+size 479860488

runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0002250.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cae5762c9f738e5d0326bd0d7bafe6e2f06d3951789dec121668f784b6012652
+size 479860488

runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0002500.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d6d24c4ec52cefe0c50e190cc93d596b3c2225536eaeea20bf2d5117883b3a74
+size 479860488

runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0002750.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:df968278694544bf7b87bdc7715a16d38858587c046b1236ec21566ca7a4b454
+size 479860488

runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0003000.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cbf258b6a6c6a6d76c47559731ee51dd24694e76fd2107f88c5923fd5c4af24b
+size 479860488

runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0003052.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bc5b1105a51cc032ea3f4272f147ea6569470c3b4dcb7d111b3d3974a67758fa
+size 479860488

runs/big_run/exp_c/run_20260409_194841/config.json

@@ -0,0 +1,91 @@
+{
+  "preprocess": null,
+  "streaming_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": 0.01287921140461293,
+    "rope_theta": 10000.0,
+    "gradient_checkpointing": false
+  },
+  "training": {
+    "device": "auto",
+    "train_bin": "data/exp_c/train.bin",
+    "val_bin": "data/exp_c/val.bin",
+    "num_workers": 2,
+    "pin_memory": true,
+    "batch_size": 32,
+    "gradient_accumulation_steps": 1,
+    "train_samples_per_epoch": null,
+    "val_sequence_stride": null,
+    "learning_rate": 0.0026291804804649314,
+    "min_lr": 0.0002629180480464931,
+    "weight_decay": 0.13113982116099054,
+    "beta1": 0.9,
+    "beta2": 0.9606057712410854,
+    "grad_clip": 0.9147944011374611,
+    "warmup_steps": 400,
+    "max_steps": 3052,
+    "lr_schedule": "wsd",
+    "lr_decay_ratio": 0.25826046407623926,
+    "z_loss_coeff": 0.00013887413811122497,
+    "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",
+        "stride": null
+      },
+      {
+        "name": "exp_c_val",
+        "path": "data/exp_c/val.bin",
+        "subset": null,
+        "split": null,
+        "stride": null
+      }
+    ]
+  },
+  "inference": null,
+  "chat": null
+}

runs/big_run/exp_c/run_20260409_194841/train.log

@@ -0,0 +1,713 @@
+2026-04-09 19:48:42 - INFO - parrotllm - Logging initialised -> runs/big_run/exp_c/run_20260409_194841/train.log
+2026-04-09 19:48:42 - INFO - parrotllm.training - device=cuda:0 | rank=0 | world_size=8 | distributed=yes
+2026-04-09 19:48:42 - INFO - parrotllm.training - checkpoint directory=/home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints | keep_last=10 | keep_best=10
+2026-04-09 19:48:42 - INFO - parrotllm.training - train windows=random-overlapping | tokens=508793600 | valid_starts=508792576 | sampled_windows_per_epoch=508792576
+2026-04-09 19:48:42 - INFO - parrotllm.training - val windows=sequential | tokens=5138325 | stride=512 | windows=10034
+2026-04-09 19:48:43 - INFO - parrotllm.training - compiling model with torch.compile...
+2026-04-09 19:48:44 - 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: 0.01287921140461293
+  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
+============================================================================================================================================
+============================================================
+2026-04-09 19:48:44 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:48:51 - INFO - parrotllm.training -   Initial val: loss=10.8930, ppl=53798.44
+2026-04-09 19:48:58 - INFO - parrotllm.training - 
+============================================================
+Starting training...
+  Steps per epoch (approx): 1987471
+  Max steps: 3052
+============================================================
+2026-04-09 19:48:58 - INFO - parrotllm.training - LR schedule=wsd | decay_ratio=0.25826046407623926 | z_loss_coeff=0.00013887413811122497
+2026-04-09 19:49:16 - INFO - parrotllm.training - step      1 | epoch 0 | loss 10.9091 | lr 6.57e-06 | grad 7.1365
+2026-04-09 19:49:19 - INFO - parrotllm.training - step     10 | epoch 0 | loss 10.1779 | lr 6.57e-05 | grad 1.7449
+2026-04-09 19:49:22 - INFO - parrotllm.training - step     20 | epoch 0 | loss 9.6632 | lr 1.31e-04 | grad 1.7042
+2026-04-09 19:49:25 - INFO - parrotllm.training - step     30 | epoch 0 | loss 8.9701 | lr 1.97e-04 | grad 1.6799
+2026-04-09 19:49:28 - INFO - parrotllm.training - step     40 | epoch 0 | loss 8.1646 | lr 2.63e-04 | grad 1.4101
+2026-04-09 19:49:31 - INFO - parrotllm.training - step     50 | epoch 0 | loss 7.5411 | lr 3.29e-04 | grad 0.9232
+2026-04-09 19:49:34 - INFO - parrotllm.training - step     60 | epoch 0 | loss 7.2607 | lr 3.94e-04 | grad 0.5898
+2026-04-09 19:49:37 - INFO - parrotllm.training - step     70 | epoch 0 | loss 7.0955 | lr 4.60e-04 | grad 1.2679
+2026-04-09 19:49:40 - INFO - parrotllm.training - step     80 | epoch 0 | loss 6.7866 | lr 5.26e-04 | grad 0.7693
+2026-04-09 19:49:43 - INFO - parrotllm.training - step     90 | epoch 0 | loss 6.6007 | lr 5.92e-04 | grad 1.0560
+2026-04-09 19:49:46 - INFO - parrotllm.training - step    100 | epoch 0 | loss 6.4954 | lr 6.57e-04 | grad 0.8125
+2026-04-09 19:49:46 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:49:46 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:49:50 - INFO - parrotllm.training -   Train: loss=6.4954, ppl=662.08
+2026-04-09 19:49:50 - INFO - parrotllm.training -   Val:   loss=6.5489, ppl=698.49
+2026-04-09 19:49:50 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:49:51 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_6p5489_epoch_0000_step_0000100.pt
+2026-04-09 19:49:52 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:49:55 - INFO - parrotllm.training - step    110 | epoch 0 | loss 6.4405 | lr 7.23e-04 | grad 0.9682
+2026-04-09 19:49:58 - INFO - parrotllm.training - step    120 | epoch 0 | loss 6.2927 | lr 7.89e-04 | grad 0.7152
+2026-04-09 19:50:01 - INFO - parrotllm.training - step    130 | epoch 0 | loss 6.1865 | lr 8.54e-04 | grad 0.9710
+2026-04-09 19:50:05 - INFO - parrotllm.training - step    140 | epoch 0 | loss 6.1820 | lr 9.20e-04 | grad 1.1712
+2026-04-09 19:50:08 - INFO - parrotllm.training - step    150 | epoch 0 | loss 6.1444 | lr 9.86e-04 | grad 1.0419
+2026-04-09 19:50:11 - INFO - parrotllm.training - step    160 | epoch 0 | loss 5.8670 | lr 1.05e-03 | grad 0.4032
+2026-04-09 19:50:14 - INFO - parrotllm.training - step    170 | epoch 0 | loss 5.9370 | lr 1.12e-03 | grad 0.7011
+2026-04-09 19:50:17 - INFO - parrotllm.training - step    180 | epoch 0 | loss 5.7945 | lr 1.18e-03 | grad 0.9991
+2026-04-09 19:50:20 - INFO - parrotllm.training - step    190 | epoch 0 | loss 5.8239 | lr 1.25e-03 | grad 1.1057
+2026-04-09 19:50:23 - INFO - parrotllm.training - step    200 | epoch 0 | loss 5.7555 | lr 1.31e-03 | grad 0.4688
+2026-04-09 19:50:23 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:50:23 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:50:27 - INFO - parrotllm.training -   Train: loss=5.7555, ppl=315.92
+2026-04-09 19:50:27 - INFO - parrotllm.training -   Val:   loss=5.7234, ppl=305.96
+2026-04-09 19:50:27 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:50:28 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_5p7234_epoch_0000_step_0000200.pt
+2026-04-09 19:50:29 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:50:32 - INFO - parrotllm.training - step    210 | epoch 0 | loss 5.7288 | lr 1.38e-03 | grad 1.0780
+2026-04-09 19:50:35 - INFO - parrotllm.training - step    220 | epoch 0 | loss 5.6321 | lr 1.45e-03 | grad 0.5545
+2026-04-09 19:50:38 - INFO - parrotllm.training - step    230 | epoch 0 | loss 5.5486 | lr 1.51e-03 | grad 0.4590
+2026-04-09 19:50:41 - INFO - parrotllm.training - step    240 | epoch 0 | loss 5.5967 | lr 1.58e-03 | grad 0.7182
+2026-04-09 19:50:44 - INFO - parrotllm.training - step    250 | epoch 0 | loss 5.5082 | lr 1.64e-03 | grad 0.5399
+2026-04-09 19:50:45 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0000250.pt
+2026-04-09 19:50:49 - INFO - parrotllm.training - step    260 | epoch 0 | loss 5.3865 | lr 1.71e-03 | grad 0.4061
+2026-04-09 19:50:52 - INFO - parrotllm.training - step    270 | epoch 0 | loss 5.3564 | lr 1.77e-03 | grad 0.7046
+2026-04-09 19:50:55 - INFO - parrotllm.training - step    280 | epoch 0 | loss 5.3435 | lr 1.84e-03 | grad 0.3155
+2026-04-09 19:50:58 - INFO - parrotllm.training - step    290 | epoch 0 | loss 5.2527 | lr 1.91e-03 | grad 0.4210
+2026-04-09 19:51:01 - INFO - parrotllm.training - step    300 | epoch 0 | loss 5.1778 | lr 1.97e-03 | grad 0.2726
+2026-04-09 19:51:01 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:51:01 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:51:05 - INFO - parrotllm.training -   Train: loss=5.1778, ppl=177.30
+2026-04-09 19:51:05 - INFO - parrotllm.training -   Val:   loss=5.1796, ppl=177.62
+2026-04-09 19:51:05 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:51:06 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_5p1796_epoch_0000_step_0000300.pt
+2026-04-09 19:51:07 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:51:10 - INFO - parrotllm.training - step    310 | epoch 0 | loss 5.1673 | lr 2.04e-03 | grad 0.4610
+2026-04-09 19:51:13 - INFO - parrotllm.training - step    320 | epoch 0 | loss 5.1955 | lr 2.10e-03 | grad 0.2374
+2026-04-09 19:51:16 - INFO - parrotllm.training - step    330 | epoch 0 | loss 5.1380 | lr 2.17e-03 | grad 0.5871
+2026-04-09 19:51:19 - INFO - parrotllm.training - step    340 | epoch 0 | loss 4.9124 | lr 2.23e-03 | grad 0.2570
+2026-04-09 19:51:22 - INFO - parrotllm.training - step    350 | epoch 0 | loss 5.0496 | lr 2.30e-03 | grad 0.4649
+2026-04-09 19:51:25 - INFO - parrotllm.training - step    360 | epoch 0 | loss 5.0396 | lr 2.37e-03 | grad 0.2588
+2026-04-09 19:51:28 - INFO - parrotllm.training - step    370 | epoch 0 | loss 5.0648 | lr 2.43e-03 | grad 0.3566
+2026-04-09 19:51:31 - INFO - parrotllm.training - step    380 | epoch 0 | loss 4.8828 | lr 2.50e-03 | grad 0.4501
+2026-04-09 19:51:34 - INFO - parrotllm.training - step    390 | epoch 0 | loss 4.7993 | lr 2.56e-03 | grad 0.2795
+2026-04-09 19:51:37 - INFO - parrotllm.training - step    400 | epoch 0 | loss 4.8585 | lr 2.63e-03 | grad 0.2891
+2026-04-09 19:51:37 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:51:37 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:51:41 - INFO - parrotllm.training -   Train: loss=4.8585, ppl=128.83
+2026-04-09 19:51:41 - INFO - parrotllm.training -   Val:   loss=4.7745, ppl=118.45
+2026-04-09 19:51:41 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:51:42 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_4p7745_epoch_0000_step_0000400.pt
+2026-04-09 19:51:43 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:51:46 - INFO - parrotllm.training - step    410 | epoch 0 | loss 4.7663 | lr 2.63e-03 | grad 0.2583
+2026-04-09 19:51:49 - INFO - parrotllm.training - step    420 | epoch 0 | loss 4.8949 | lr 2.63e-03 | grad 0.3631
+2026-04-09 19:51:52 - INFO - parrotllm.training - step    430 | epoch 0 | loss 4.7069 | lr 2.63e-03 | grad 0.3115
+2026-04-09 19:51:55 - INFO - parrotllm.training - step    440 | epoch 0 | loss 4.7118 | lr 2.63e-03 | grad 0.2092
+2026-04-09 19:51:58 - INFO - parrotllm.training - step    450 | epoch 0 | loss 4.6394 | lr 2.63e-03 | grad 0.3845
+2026-04-09 19:52:01 - INFO - parrotllm.training - step    460 | epoch 0 | loss 4.6215 | lr 2.63e-03 | grad 0.2960
+2026-04-09 19:52:04 - INFO - parrotllm.training - step    470 | epoch 0 | loss 4.5301 | lr 2.63e-03 | grad 0.2773
+2026-04-09 19:52:07 - INFO - parrotllm.training - step    480 | epoch 0 | loss 4.5227 | lr 2.63e-03 | grad 0.2836
+2026-04-09 19:52:11 - INFO - parrotllm.training - step    490 | epoch 0 | loss 4.5967 | lr 2.63e-03 | grad 0.2442
+2026-04-09 19:52:14 - INFO - parrotllm.training - step    500 | epoch 0 | loss 4.4588 | lr 2.63e-03 | grad 0.3044
+2026-04-09 19:52:14 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:52:14 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:52:18 - INFO - parrotllm.training -   Train: loss=4.4588, ppl=86.38
+2026-04-09 19:52:18 - INFO - parrotllm.training -   Val:   loss=4.4215, ppl=83.22
+2026-04-09 19:52:18 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:52:19 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_4p4215_epoch_0000_step_0000500.pt
+2026-04-09 19:52:20 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:52:21 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0000500.pt
+2026-04-09 19:52:24 - INFO - parrotllm.training - step    510 | epoch 0 | loss 4.4188 | lr 2.63e-03 | grad 0.3133
+2026-04-09 19:52:28 - INFO - parrotllm.training - step    520 | epoch 0 | loss 4.5455 | lr 2.63e-03 | grad 0.2198
+2026-04-09 19:52:31 - INFO - parrotllm.training - step    530 | epoch 0 | loss 4.4403 | lr 2.63e-03 | grad 0.2201
+2026-04-09 19:52:34 - INFO - parrotllm.training - step    540 | epoch 0 | loss 4.4742 | lr 2.63e-03 | grad 0.2426
+2026-04-09 19:52:37 - INFO - parrotllm.training - step    550 | epoch 0 | loss 4.3802 | lr 2.63e-03 | grad 0.2205
+2026-04-09 19:52:40 - INFO - parrotllm.training - step    560 | epoch 0 | loss 4.3076 | lr 2.63e-03 | grad 0.2276
+2026-04-09 19:52:43 - INFO - parrotllm.training - step    570 | epoch 0 | loss 4.3453 | lr 2.63e-03 | grad 0.2497
+2026-04-09 19:52:46 - INFO - parrotllm.training - step    580 | epoch 0 | loss 4.2113 | lr 2.63e-03 | grad 0.2698
+2026-04-09 19:52:49 - INFO - parrotllm.training - step    590 | epoch 0 | loss 4.2884 | lr 2.63e-03 | grad 0.1941
+2026-04-09 19:52:52 - INFO - parrotllm.training - step    600 | epoch 0 | loss 4.3145 | lr 2.63e-03 | grad 0.2591
+2026-04-09 19:52:52 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:52:52 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:52:56 - INFO - parrotllm.training -   Train: loss=4.3145, ppl=74.78
+2026-04-09 19:52:56 - INFO - parrotllm.training -   Val:   loss=4.2254, ppl=68.40
+2026-04-09 19:52:56 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:52:57 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_4p2254_epoch_0000_step_0000600.pt
+2026-04-09 19:52:58 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:53:01 - INFO - parrotllm.training - step    610 | epoch 0 | loss 4.2772 | lr 2.63e-03 | grad 0.2242
+2026-04-09 19:53:04 - INFO - parrotllm.training - step    620 | epoch 0 | loss 4.3137 | lr 2.63e-03 | grad 0.2324
+2026-04-09 19:53:07 - INFO - parrotllm.training - step    630 | epoch 0 | loss 4.2215 | lr 2.63e-03 | grad 0.2176
+2026-04-09 19:53:10 - INFO - parrotllm.training - step    640 | epoch 0 | loss 4.2153 | lr 2.63e-03 | grad 0.2100
+2026-04-09 19:53:13 - INFO - parrotllm.training - step    650 | epoch 0 | loss 4.2877 | lr 2.63e-03 | grad 0.2277
+2026-04-09 19:53:16 - INFO - parrotllm.training - step    660 | epoch 0 | loss 4.1158 | lr 2.63e-03 | grad 0.2027
+2026-04-09 19:53:19 - INFO - parrotllm.training - step    670 | epoch 0 | loss 4.2387 | lr 2.63e-03 | grad 0.2492
+2026-04-09 19:53:22 - INFO - parrotllm.training - step    680 | epoch 0 | loss 4.2427 | lr 2.63e-03 | grad 0.2030
+2026-04-09 19:53:25 - INFO - parrotllm.training - step    690 | epoch 0 | loss 4.0899 | lr 2.63e-03 | grad 0.2669
+2026-04-09 19:53:28 - INFO - parrotllm.training - step    700 | epoch 0 | loss 4.1790 | lr 2.63e-03 | grad 0.1845
+2026-04-09 19:53:28 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:53:28 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:53:32 - INFO - parrotllm.training -   Train: loss=4.1790, ppl=65.30
+2026-04-09 19:53:32 - INFO - parrotllm.training -   Val:   loss=4.1070, ppl=60.76
+2026-04-09 19:53:32 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:53:33 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_4p1070_epoch_0000_step_0000700.pt
+2026-04-09 19:53:34 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:53:37 - INFO - parrotllm.training - step    710 | epoch 0 | loss 4.1139 | lr 2.63e-03 | grad 0.2054
+2026-04-09 19:53:40 - INFO - parrotllm.training - step    720 | epoch 0 | loss 4.1864 | lr 2.63e-03 | grad 0.1854
+2026-04-09 19:53:43 - INFO - parrotllm.training - step    730 | epoch 0 | loss 4.0474 | lr 2.63e-03 | grad 0.1847
+2026-04-09 19:53:46 - INFO - parrotllm.training - step    740 | epoch 0 | loss 4.0799 | lr 2.63e-03 | grad 0.2173
+2026-04-09 19:53:49 - INFO - parrotllm.training - step    750 | epoch 0 | loss 4.0885 | lr 2.63e-03 | grad 0.2270
+2026-04-09 19:53:50 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0000750.pt
+2026-04-09 19:53:54 - INFO - parrotllm.training - step    760 | epoch 0 | loss 4.1032 | lr 2.63e-03 | grad 0.1718
+2026-04-09 19:53:57 - INFO - parrotllm.training - step    770 | epoch 0 | loss 4.0238 | lr 2.63e-03 | grad 0.2184
+2026-04-09 19:54:00 - INFO - parrotllm.training - step    780 | epoch 0 | loss 4.1056 | lr 2.63e-03 | grad 0.1994
+2026-04-09 19:54:03 - INFO - parrotllm.training - step    790 | epoch 0 | loss 4.1240 | lr 2.63e-03 | grad 0.1723
+2026-04-09 19:54:06 - INFO - parrotllm.training - step    800 | epoch 0 | loss 4.1634 | lr 2.63e-03 | grad 0.2236
+2026-04-09 19:54:06 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:54:06 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:54:10 - INFO - parrotllm.training -   Train: loss=4.1634, ppl=64.29
+2026-04-09 19:54:10 - INFO - parrotllm.training -   Val:   loss=4.0321, ppl=56.38
+2026-04-09 19:54:10 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:54:11 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_4p0321_epoch_0000_step_0000800.pt
+2026-04-09 19:54:12 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:54:15 - INFO - parrotllm.training - step    810 | epoch 0 | loss 3.9682 | lr 2.63e-03 | grad 0.1716
+2026-04-09 19:54:18 - INFO - parrotllm.training - step    820 | epoch 0 | loss 4.0207 | lr 2.63e-03 | grad 0.2080
+2026-04-09 19:54:21 - INFO - parrotllm.training - step    830 | epoch 0 | loss 4.0948 | lr 2.63e-03 | grad 0.1806
+2026-04-09 19:54:24 - INFO - parrotllm.training - step    840 | epoch 0 | loss 4.0619 | lr 2.63e-03 | grad 0.1814
+2026-04-09 19:54:27 - INFO - parrotllm.training - step    850 | epoch 0 | loss 4.0035 | lr 2.63e-03 | grad 0.2343
+2026-04-09 19:54:30 - INFO - parrotllm.training - step    860 | epoch 0 | loss 3.9593 | lr 2.63e-03 | grad 0.1733
+2026-04-09 19:54:33 - INFO - parrotllm.training - step    870 | epoch 0 | loss 4.0034 | lr 2.63e-03 | grad 0.2086
+2026-04-09 19:54:36 - INFO - parrotllm.training - step    880 | epoch 0 | loss 4.0828 | lr 2.63e-03 | grad 0.1862
+2026-04-09 19:54:39 - INFO - parrotllm.training - step    890 | epoch 0 | loss 4.0499 | lr 2.63e-03 | grad 0.1666
+2026-04-09 19:54:42 - INFO - parrotllm.training - step    900 | epoch 0 | loss 3.9918 | lr 2.63e-03 | grad 0.1634
+2026-04-09 19:54:42 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:54:42 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:54:47 - INFO - parrotllm.training -   Train: loss=3.9918, ppl=54.15
+2026-04-09 19:54:47 - INFO - parrotllm.training -   Val:   loss=3.9588, ppl=52.40
+2026-04-09 19:54:47 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:54:47 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p9588_epoch_0000_step_0000900.pt
+2026-04-09 19:54:48 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:54:51 - INFO - parrotllm.training - step    910 | epoch 0 | loss 4.0519 | lr 2.63e-03 | grad 0.1760
+2026-04-09 19:54:54 - INFO - parrotllm.training - step    920 | epoch 0 | loss 4.0651 | lr 2.63e-03 | grad 0.2031
+2026-04-09 19:54:57 - INFO - parrotllm.training - step    930 | epoch 0 | loss 4.0251 | lr 2.63e-03 | grad 0.1630
+2026-04-09 19:55:00 - INFO - parrotllm.training - step    940 | epoch 0 | loss 4.0553 | lr 2.63e-03 | grad 0.1557
+2026-04-09 19:55:03 - INFO - parrotllm.training - step    950 | epoch 0 | loss 4.0912 | lr 2.63e-03 | grad 0.1676
+2026-04-09 19:55:06 - INFO - parrotllm.training - step    960 | epoch 0 | loss 3.9219 | lr 2.63e-03 | grad 0.1715
+2026-04-09 19:55:09 - INFO - parrotllm.training - step    970 | epoch 0 | loss 4.1238 | lr 2.63e-03 | grad 0.1910
+2026-04-09 19:55:12 - INFO - parrotllm.training - step    980 | epoch 0 | loss 3.9491 | lr 2.63e-03 | grad 0.1557
+2026-04-09 19:55:16 - INFO - parrotllm.training - step    990 | epoch 0 | loss 3.8290 | lr 2.63e-03 | grad 0.2054
+2026-04-09 19:55:19 - INFO - parrotllm.training - step   1000 | epoch 0 | loss 4.0221 | lr 2.63e-03 | grad 0.2379
+2026-04-09 19:55:19 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:55:19 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:55:23 - INFO - parrotllm.training -   Train: loss=4.0221, ppl=55.82
+2026-04-09 19:55:23 - INFO - parrotllm.training -   Val:   loss=3.9146, ppl=50.13
+2026-04-09 19:55:23 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:55:24 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p9146_epoch_0000_step_0001000.pt
+2026-04-09 19:55:24 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:55:25 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0001000.pt
+2026-04-09 19:55:29 - INFO - parrotllm.training - step   1010 | epoch 0 | loss 3.8819 | lr 2.63e-03 | grad 0.1976
+2026-04-09 19:55:32 - INFO - parrotllm.training - step   1020 | epoch 0 | loss 3.9002 | lr 2.63e-03 | grad 0.1956
+2026-04-09 19:55:35 - INFO - parrotllm.training - step   1030 | epoch 0 | loss 3.9900 | lr 2.63e-03 | grad 0.2064
+2026-04-09 19:55:38 - INFO - parrotllm.training - step   1040 | epoch 0 | loss 4.0041 | lr 2.63e-03 | grad 0.1579
+2026-04-09 19:55:41 - INFO - parrotllm.training - step   1050 | epoch 0 | loss 3.9342 | lr 2.63e-03 | grad 0.1911
+2026-04-09 19:55:44 - INFO - parrotllm.training - step   1060 | epoch 0 | loss 3.9709 | lr 2.63e-03 | grad 0.1651
+2026-04-09 19:55:47 - INFO - parrotllm.training - step   1070 | epoch 0 | loss 4.0302 | lr 2.63e-03 | grad 0.2034
+2026-04-09 19:55:50 - INFO - parrotllm.training - step   1080 | epoch 0 | loss 4.0819 | lr 2.63e-03 | grad 0.1742
+2026-04-09 19:55:53 - INFO - parrotllm.training - step   1090 | epoch 0 | loss 3.9620 | lr 2.63e-03 | grad 0.1613
+2026-04-09 19:55:56 - INFO - parrotllm.training - step   1100 | epoch 0 | loss 3.9085 | lr 2.63e-03 | grad 0.2163
+2026-04-09 19:55:56 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:55:56 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:56:01 - INFO - parrotllm.training -   Train: loss=3.9085, ppl=49.82
+2026-04-09 19:56:01 - INFO - parrotllm.training -   Val:   loss=3.8782, ppl=48.34
+2026-04-09 19:56:01 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:56:02 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p8782_epoch_0000_step_0001100.pt
+2026-04-09 19:56:02 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:56:05 - INFO - parrotllm.training - step   1110 | epoch 0 | loss 3.8925 | lr 2.63e-03 | grad 0.1699
+2026-04-09 19:56:08 - INFO - parrotllm.training - step   1120 | epoch 0 | loss 3.9041 | lr 2.63e-03 | grad 0.1541
+2026-04-09 19:56:11 - INFO - parrotllm.training - step   1130 | epoch 0 | loss 3.9580 | lr 2.63e-03 | grad 0.2096
+2026-04-09 19:56:14 - INFO - parrotllm.training - step   1140 | epoch 0 | loss 3.9132 | lr 2.63e-03 | grad 0.2008
+2026-04-09 19:56:17 - INFO - parrotllm.training - step   1150 | epoch 0 | loss 3.9540 | lr 2.63e-03 | grad 0.1475
+2026-04-09 19:56:20 - INFO - parrotllm.training - step   1160 | epoch 0 | loss 3.8995 | lr 2.63e-03 | grad 0.1653
+2026-04-09 19:56:23 - INFO - parrotllm.training - step   1170 | epoch 0 | loss 3.9661 | lr 2.63e-03 | grad 0.1744
+2026-04-09 19:56:27 - INFO - parrotllm.training - step   1180 | epoch 0 | loss 3.8537 | lr 2.63e-03 | grad 0.1642
+2026-04-09 19:56:30 - INFO - parrotllm.training - step   1190 | epoch 0 | loss 3.8895 | lr 2.63e-03 | grad 0.1736
+2026-04-09 19:56:33 - INFO - parrotllm.training - step   1200 | epoch 0 | loss 3.9463 | lr 2.63e-03 | grad 0.1488
+2026-04-09 19:56:33 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:56:33 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:56:37 - INFO - parrotllm.training -   Train: loss=3.9463, ppl=51.74
+2026-04-09 19:56:37 - INFO - parrotllm.training -   Val:   loss=3.8392, ppl=46.49
+2026-04-09 19:56:37 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:56:38 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p8392_epoch_0000_step_0001200.pt
+2026-04-09 19:56:39 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:56:42 - INFO - parrotllm.training - step   1210 | epoch 0 | loss 3.9002 | lr 2.63e-03 | grad 0.1699
+2026-04-09 19:56:45 - INFO - parrotllm.training - step   1220 | epoch 0 | loss 3.9044 | lr 2.63e-03 | grad 0.1731
+2026-04-09 19:56:48 - INFO - parrotllm.training - step   1230 | epoch 0 | loss 3.8746 | lr 2.63e-03 | grad 0.1867
+2026-04-09 19:56:51 - INFO - parrotllm.training - step   1240 | epoch 0 | loss 3.8094 | lr 2.63e-03 | grad 0.1599
+2026-04-09 19:56:54 - INFO - parrotllm.training - step   1250 | epoch 0 | loss 3.8403 | lr 2.63e-03 | grad 0.1822
+2026-04-09 19:56:55 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0001250.pt
+2026-04-09 19:56:59 - INFO - parrotllm.training - step   1260 | epoch 0 | loss 3.6994 | lr 2.63e-03 | grad 0.2033
+2026-04-09 19:57:02 - INFO - parrotllm.training - step   1270 | epoch 0 | loss 3.8480 | lr 2.63e-03 | grad 0.1652
+2026-04-09 19:57:05 - INFO - parrotllm.training - step   1280 | epoch 0 | loss 3.8095 | lr 2.63e-03 | grad 0.1429
+2026-04-09 19:57:08 - INFO - parrotllm.training - step   1290 | epoch 0 | loss 3.9075 | lr 2.63e-03 | grad 0.1846
+2026-04-09 19:57:11 - INFO - parrotllm.training - step   1300 | epoch 0 | loss 3.7979 | lr 2.63e-03 | grad 0.1781
+2026-04-09 19:57:11 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:57:11 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:57:15 - INFO - parrotllm.training -   Train: loss=3.7979, ppl=44.61
+2026-04-09 19:57:15 - INFO - parrotllm.training -   Val:   loss=3.8142, ppl=45.34
+2026-04-09 19:57:15 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:57:16 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p8142_epoch_0000_step_0001300.pt
+2026-04-09 19:57:17 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:57:20 - INFO - parrotllm.training - step   1310 | epoch 0 | loss 3.8241 | lr 2.63e-03 | grad 0.1577
+2026-04-09 19:57:23 - INFO - parrotllm.training - step   1320 | epoch 0 | loss 3.8700 | lr 2.63e-03 | grad 0.1788
+2026-04-09 19:57:26 - INFO - parrotllm.training - step   1330 | epoch 0 | loss 3.7728 | lr 2.63e-03 | grad 0.1649
+2026-04-09 19:57:29 - INFO - parrotllm.training - step   1340 | epoch 0 | loss 3.9151 | lr 2.63e-03 | grad 0.1587
+2026-04-09 19:57:32 - INFO - parrotllm.training - step   1350 | epoch 0 | loss 3.7757 | lr 2.63e-03 | grad 0.1523
+2026-04-09 19:57:35 - INFO - parrotllm.training - step   1360 | epoch 0 | loss 3.8458 | lr 2.63e-03 | grad 0.1831
+2026-04-09 19:57:38 - INFO - parrotllm.training - step   1370 | epoch 0 | loss 3.9368 | lr 2.63e-03 | grad 0.1795
+2026-04-09 19:57:41 - INFO - parrotllm.training - step   1380 | epoch 0 | loss 3.7332 | lr 2.63e-03 | grad 0.1437
+2026-04-09 19:57:44 - INFO - parrotllm.training - step   1390 | epoch 0 | loss 3.9803 | lr 2.63e-03 | grad 0.1775
+2026-04-09 19:57:47 - INFO - parrotllm.training - step   1400 | epoch 0 | loss 3.7294 | lr 2.63e-03 | grad 0.1528
+2026-04-09 19:57:47 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:57:47 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:57:51 - INFO - parrotllm.training -   Train: loss=3.7294, ppl=41.65
+2026-04-09 19:57:51 - INFO - parrotllm.training -   Val:   loss=3.7858, ppl=44.07
+2026-04-09 19:57:51 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:57:52 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p7858_epoch_0000_step_0001400.pt
+2026-04-09 19:57:53 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:57:56 - INFO - parrotllm.training - step   1410 | epoch 0 | loss 3.7361 | lr 2.63e-03 | grad 0.1693
+2026-04-09 19:57:59 - INFO - parrotllm.training - step   1420 | epoch 0 | loss 3.8784 | lr 2.63e-03 | grad 0.1485
+2026-04-09 19:58:02 - INFO - parrotllm.training - step   1430 | epoch 0 | loss 3.8316 | lr 2.63e-03 | grad 0.1922
+2026-04-09 19:58:05 - INFO - parrotllm.training - step   1440 | epoch 0 | loss 3.8551 | lr 2.63e-03 | grad 0.1819
+2026-04-09 19:58:08 - INFO - parrotllm.training - step   1450 | epoch 0 | loss 3.8035 | lr 2.63e-03 | grad 0.1487
+2026-04-09 19:58:11 - INFO - parrotllm.training - step   1460 | epoch 0 | loss 3.9258 | lr 2.63e-03 | grad 0.1667
+2026-04-09 19:58:15 - INFO - parrotllm.training - step   1470 | epoch 0 | loss 3.8732 | lr 2.63e-03 | grad 0.2148
+2026-04-09 19:58:18 - INFO - parrotllm.training - step   1480 | epoch 0 | loss 3.7913 | lr 2.63e-03 | grad 0.1656
+2026-04-09 19:58:21 - INFO - parrotllm.training - step   1490 | epoch 0 | loss 3.7830 | lr 2.63e-03 | grad 0.1748
+2026-04-09 19:58:24 - INFO - parrotllm.training - step   1500 | epoch 0 | loss 3.7754 | lr 2.63e-03 | grad 0.1659
+2026-04-09 19:58:24 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:58:24 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:58:28 - INFO - parrotllm.training -   Train: loss=3.7754, ppl=43.61
+2026-04-09 19:58:28 - INFO - parrotllm.training -   Val:   loss=3.7697, ppl=43.37
+2026-04-09 19:58:28 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:58:29 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p7697_epoch_0000_step_0001500.pt
+2026-04-09 19:58:30 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:58:31 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0001500.pt
+2026-04-09 19:58:35 - INFO - parrotllm.training - step   1510 | epoch 0 | loss 3.8747 | lr 2.63e-03 | grad 0.1409
+2026-04-09 19:58:38 - INFO - parrotllm.training - step   1520 | epoch 0 | loss 3.7320 | lr 2.63e-03 | grad 0.1543
+2026-04-09 19:58:41 - INFO - parrotllm.training - step   1530 | epoch 0 | loss 3.8278 | lr 2.63e-03 | grad 0.1704
+2026-04-09 19:58:44 - INFO - parrotllm.training - step   1540 | epoch 0 | loss 3.7961 | lr 2.63e-03 | grad 0.1458
+2026-04-09 19:58:47 - INFO - parrotllm.training - step   1550 | epoch 0 | loss 3.8211 | lr 2.63e-03 | grad 0.1955
+2026-04-09 19:58:50 - INFO - parrotllm.training - step   1560 | epoch 0 | loss 3.8359 | lr 2.63e-03 | grad 0.1429
+2026-04-09 19:58:53 - INFO - parrotllm.training - step   1570 | epoch 0 | loss 3.8386 | lr 2.63e-03 | grad 0.1450
+2026-04-09 19:58:56 - INFO - parrotllm.training - step   1580 | epoch 0 | loss 3.7798 | lr 2.63e-03 | grad 0.1654
+2026-04-09 19:58:59 - INFO - parrotllm.training - step   1590 | epoch 0 | loss 3.7744 | lr 2.63e-03 | grad 0.2094
+2026-04-09 19:59:02 - INFO - parrotllm.training - step   1600 | epoch 0 | loss 3.8904 | lr 2.63e-03 | grad 0.1696
+2026-04-09 19:59:02 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:59:02 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:59:07 - INFO - parrotllm.training -   Train: loss=3.8904, ppl=48.93
+2026-04-09 19:59:07 - INFO - parrotllm.training -   Val:   loss=3.7493, ppl=42.49
+2026-04-09 19:59:07 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:59:08 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p7493_epoch_0000_step_0001600.pt
+2026-04-09 19:59:08 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:59:11 - INFO - parrotllm.training - step   1610 | epoch 0 | loss 3.7062 | lr 2.63e-03 | grad 0.1427
+2026-04-09 19:59:14 - INFO - parrotllm.training - step   1620 | epoch 0 | loss 3.7471 | lr 2.63e-03 | grad 0.1440
+2026-04-09 19:59:17 - INFO - parrotllm.training - step   1630 | epoch 0 | loss 3.8186 | lr 2.63e-03 | grad 0.1761
+2026-04-09 19:59:21 - INFO - parrotllm.training - step   1640 | epoch 0 | loss 3.7734 | lr 2.63e-03 | grad 0.1571
+2026-04-09 19:59:24 - INFO - parrotllm.training - step   1650 | epoch 0 | loss 3.7052 | lr 2.63e-03 | grad 0.1930
+2026-04-09 19:59:27 - INFO - parrotllm.training - step   1660 | epoch 0 | loss 3.8150 | lr 2.63e-03 | grad 0.1619
+2026-04-09 19:59:30 - INFO - parrotllm.training - step   1670 | epoch 0 | loss 3.9116 | lr 2.63e-03 | grad 0.1553
+2026-04-09 19:59:33 - INFO - parrotllm.training - step   1680 | epoch 0 | loss 3.8220 | lr 2.63e-03 | grad 0.1567
+2026-04-09 19:59:36 - INFO - parrotllm.training - step   1690 | epoch 0 | loss 3.7747 | lr 2.63e-03 | grad 0.1546
+2026-04-09 19:59:39 - INFO - parrotllm.training - step   1700 | epoch 0 | loss 3.8073 | lr 2.63e-03 | grad 0.1757
+2026-04-09 19:59:39 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 19:59:39 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:59:43 - INFO - parrotllm.training -   Train: loss=3.8073, ppl=45.03
+2026-04-09 19:59:43 - INFO - parrotllm.training -   Val:   loss=3.7333, ppl=41.82
+2026-04-09 19:59:43 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 19:59:44 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p7333_epoch_0000_step_0001700.pt
+2026-04-09 19:59:45 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 19:59:48 - INFO - parrotllm.training - step   1710 | epoch 0 | loss 3.8052 | lr 2.63e-03 | grad 0.1558
+2026-04-09 19:59:51 - INFO - parrotllm.training - step   1720 | epoch 0 | loss 3.7411 | lr 2.63e-03 | grad 0.1648
+2026-04-09 19:59:54 - INFO - parrotllm.training - step   1730 | epoch 0 | loss 3.7603 | lr 2.63e-03 | grad 0.1408
+2026-04-09 19:59:57 - INFO - parrotllm.training - step   1740 | epoch 0 | loss 3.7127 | lr 2.63e-03 | grad 0.1539
+2026-04-09 20:00:00 - INFO - parrotllm.training - step   1750 | epoch 0 | loss 3.7317 | lr 2.63e-03 | grad 0.1736
+2026-04-09 20:00:01 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0001750.pt
+2026-04-09 20:00:05 - INFO - parrotllm.training - step   1760 | epoch 0 | loss 3.9323 | lr 2.63e-03 | grad 0.1407
+2026-04-09 20:00:08 - INFO - parrotllm.training - step   1770 | epoch 0 | loss 3.7911 | lr 2.63e-03 | grad 0.1666
+2026-04-09 20:00:11 - INFO - parrotllm.training - step   1780 | epoch 0 | loss 3.6901 | lr 2.63e-03 | grad 0.1531
+2026-04-09 20:00:14 - INFO - parrotllm.training - step   1790 | epoch 0 | loss 3.7748 | lr 2.63e-03 | grad 0.1655
+2026-04-09 20:00:17 - INFO - parrotllm.training - step   1800 | epoch 0 | loss 3.7228 | lr 2.63e-03 | grad 0.1427
+2026-04-09 20:00:17 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:00:17 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:00:21 - INFO - parrotllm.training -   Train: loss=3.7228, ppl=41.38
+2026-04-09 20:00:21 - INFO - parrotllm.training -   Val:   loss=3.7169, ppl=41.14
+2026-04-09 20:00:21 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:00:22 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p7169_epoch_0000_step_0001800.pt
+2026-04-09 20:00:23 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:00:26 - INFO - parrotllm.training - step   1810 | epoch 0 | loss 3.8436 | lr 2.63e-03 | grad 0.1847
+2026-04-09 20:00:29 - INFO - parrotllm.training - step   1820 | epoch 0 | loss 3.7631 | lr 2.63e-03 | grad 0.1814
+2026-04-09 20:00:32 - INFO - parrotllm.training - step   1830 | epoch 0 | loss 3.7805 | lr 2.63e-03 | grad 0.1395
+2026-04-09 20:00:35 - INFO - parrotllm.training - step   1840 | epoch 0 | loss 3.7153 | lr 2.63e-03 | grad 0.1972
+2026-04-09 20:00:38 - INFO - parrotllm.training - step   1850 | epoch 0 | loss 3.7044 | lr 2.63e-03 | grad 0.1446
+2026-04-09 20:00:41 - INFO - parrotllm.training - step   1860 | epoch 0 | loss 3.6801 | lr 2.63e-03 | grad 0.1872
+2026-04-09 20:00:44 - INFO - parrotllm.training - step   1870 | epoch 0 | loss 3.8932 | lr 2.63e-03 | grad 0.1416
+2026-04-09 20:00:47 - INFO - parrotllm.training - step   1880 | epoch 0 | loss 3.8214 | lr 2.63e-03 | grad 0.1432
+2026-04-09 20:00:50 - INFO - parrotllm.training - step   1890 | epoch 0 | loss 3.7863 | lr 2.63e-03 | grad 0.1723
+2026-04-09 20:00:53 - INFO - parrotllm.training - step   1900 | epoch 0 | loss 3.7739 | lr 2.63e-03 | grad 0.1862
+2026-04-09 20:00:53 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:00:53 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:00:58 - INFO - parrotllm.training -   Train: loss=3.7739, ppl=43.55
+2026-04-09 20:00:58 - INFO - parrotllm.training -   Val:   loss=3.7052, ppl=40.66
+2026-04-09 20:00:58 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:00:59 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p7052_epoch_0000_step_0001900.pt
+2026-04-09 20:00:59 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:01:02 - INFO - parrotllm.training - step   1910 | epoch 0 | loss 3.7373 | lr 2.63e-03 | grad 0.1450
+2026-04-09 20:01:05 - INFO - parrotllm.training - step   1920 | epoch 0 | loss 3.8460 | lr 2.63e-03 | grad 0.1306
+2026-04-09 20:01:08 - INFO - parrotllm.training - step   1930 | epoch 0 | loss 3.7930 | lr 2.63e-03 | grad 0.1486
+2026-04-09 20:01:11 - INFO - parrotllm.training - step   1940 | epoch 0 | loss 3.7708 | lr 2.63e-03 | grad 0.1497
+2026-04-09 20:01:15 - INFO - parrotllm.training - step   1950 | epoch 0 | loss 3.7059 | lr 2.63e-03 | grad 0.1552
+2026-04-09 20:01:18 - INFO - parrotllm.training - step   1960 | epoch 0 | loss 3.7852 | lr 2.63e-03 | grad 0.1554
+2026-04-09 20:01:21 - INFO - parrotllm.training - step   1970 | epoch 0 | loss 3.6680 | lr 2.63e-03 | grad 0.2033
+2026-04-09 20:01:24 - INFO - parrotllm.training - step   1980 | epoch 0 | loss 3.7221 | lr 2.63e-03 | grad 0.1510
+2026-04-09 20:01:27 - INFO - parrotllm.training - step   1990 | epoch 0 | loss 3.7439 | lr 2.63e-03 | grad 0.1636
+2026-04-09 20:01:30 - INFO - parrotllm.training - step   2000 | epoch 0 | loss 3.7827 | lr 2.63e-03 | grad 0.1812
+2026-04-09 20:01:30 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:01:30 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:01:34 - INFO - parrotllm.training -   Train: loss=3.7827, ppl=43.94
+2026-04-09 20:01:34 - INFO - parrotllm.training -   Val:   loss=3.6953, ppl=40.26
+2026-04-09 20:01:34 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:01:35 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p6953_epoch_0000_step_0002000.pt
+2026-04-09 20:01:36 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:01:37 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0002000.pt
+2026-04-09 20:01:41 - INFO - parrotllm.training - step   2010 | epoch 0 | loss 3.7378 | lr 2.63e-03 | grad 0.1778
+2026-04-09 20:01:44 - INFO - parrotllm.training - step   2020 | epoch 0 | loss 3.7661 | lr 2.63e-03 | grad 0.1819
+2026-04-09 20:01:47 - INFO - parrotllm.training - step   2030 | epoch 0 | loss 3.8263 | lr 2.63e-03 | grad 0.1585
+2026-04-09 20:01:50 - INFO - parrotllm.training - step   2040 | epoch 0 | loss 3.6743 | lr 2.63e-03 | grad 0.1553
+2026-04-09 20:01:53 - INFO - parrotllm.training - step   2050 | epoch 0 | loss 3.6095 | lr 2.63e-03 | grad 0.1902
+2026-04-09 20:01:56 - INFO - parrotllm.training - step   2060 | epoch 0 | loss 3.7855 | lr 2.63e-03 | grad 0.1463
+2026-04-09 20:01:59 - INFO - parrotllm.training - step   2070 | epoch 0 | loss 3.7186 | lr 2.63e-03 | grad 0.1528
+2026-04-09 20:02:02 - INFO - parrotllm.training - step   2080 | epoch 0 | loss 3.7021 | lr 2.63e-03 | grad 0.1660
+2026-04-09 20:02:05 - INFO - parrotllm.training - step   2090 | epoch 0 | loss 3.7493 | lr 2.63e-03 | grad 0.1430
+2026-04-09 20:02:08 - INFO - parrotllm.training - step   2100 | epoch 0 | loss 3.7570 | lr 2.63e-03 | grad 0.1586
+2026-04-09 20:02:08 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:02:08 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:02:12 - INFO - parrotllm.training -   Train: loss=3.7570, ppl=42.82
+2026-04-09 20:02:12 - INFO - parrotllm.training -   Val:   loss=3.6834, ppl=39.78
+2026-04-09 20:02:12 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:02:14 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p6834_epoch_0000_step_0002100.pt
+2026-04-09 20:02:14 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:02:17 - INFO - parrotllm.training - step   2110 | epoch 0 | loss 3.7110 | lr 2.63e-03 | grad 0.1489
+2026-04-09 20:02:20 - INFO - parrotllm.training - step   2120 | epoch 0 | loss 3.7509 | lr 2.63e-03 | grad 0.1363
+2026-04-09 20:02:23 - INFO - parrotllm.training - step   2130 | epoch 0 | loss 3.7164 | lr 2.63e-03 | grad 0.1779
+2026-04-09 20:02:26 - INFO - parrotllm.training - step   2140 | epoch 0 | loss 3.7504 | lr 2.63e-03 | grad 0.1539
+2026-04-09 20:02:30 - INFO - parrotllm.training - step   2150 | epoch 0 | loss 3.7992 | lr 2.63e-03 | grad 0.1636
+2026-04-09 20:02:33 - INFO - parrotllm.training - step   2160 | epoch 0 | loss 3.6635 | lr 2.63e-03 | grad 0.1784
+2026-04-09 20:02:36 - INFO - parrotllm.training - step   2170 | epoch 0 | loss 3.6552 | lr 2.63e-03 | grad 0.1509
+2026-04-09 20:02:39 - INFO - parrotllm.training - step   2180 | epoch 0 | loss 3.7179 | lr 2.63e-03 | grad 0.1409
+2026-04-09 20:02:42 - INFO - parrotllm.training - step   2190 | epoch 0 | loss 3.7302 | lr 2.63e-03 | grad 0.1595
+2026-04-09 20:02:45 - INFO - parrotllm.training - step   2200 | epoch 0 | loss 3.7148 | lr 2.63e-03 | grad 0.1664
+2026-04-09 20:02:45 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:02:45 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:02:49 - INFO - parrotllm.training -   Train: loss=3.7148, ppl=41.05
+2026-04-09 20:02:49 - INFO - parrotllm.training -   Val:   loss=3.6726, ppl=39.35
+2026-04-09 20:02:49 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:02:50 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p6726_epoch_0000_step_0002200.pt
+2026-04-09 20:02:51 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:02:54 - INFO - parrotllm.training - step   2210 | epoch 0 | loss 3.6857 | lr 2.63e-03 | grad 0.1343
+2026-04-09 20:02:57 - INFO - parrotllm.training - step   2220 | epoch 0 | loss 3.7386 | lr 2.63e-03 | grad 0.1627
+2026-04-09 20:03:00 - INFO - parrotllm.training - step   2230 | epoch 0 | loss 3.6619 | lr 2.63e-03 | grad 0.1597
+2026-04-09 20:03:03 - INFO - parrotllm.training - step   2240 | epoch 0 | loss 3.6317 | lr 2.63e-03 | grad 0.2137
+2026-04-09 20:03:06 - INFO - parrotllm.training - step   2250 | epoch 0 | loss 3.6500 | lr 2.63e-03 | grad 0.1504
+2026-04-09 20:03:07 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0002250.pt
+2026-04-09 20:03:11 - INFO - parrotllm.training - step   2260 | epoch 0 | loss 3.7816 | lr 2.63e-03 | grad 0.1364
+2026-04-09 20:03:14 - INFO - parrotllm.training - step   2270 | epoch 0 | loss 3.6496 | lr 2.61e-03 | grad 0.1484
+2026-04-09 20:03:17 - INFO - parrotllm.training - step   2280 | epoch 0 | loss 3.7601 | lr 2.58e-03 | grad 0.1470
+2026-04-09 20:03:20 - INFO - parrotllm.training - step   2290 | epoch 0 | loss 3.8082 | lr 2.55e-03 | grad 0.1598
+2026-04-09 20:03:23 - INFO - parrotllm.training - step   2300 | epoch 0 | loss 3.7156 | lr 2.52e-03 | grad 0.1485
+2026-04-09 20:03:23 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:03:23 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:03:27 - INFO - parrotllm.training -   Train: loss=3.7156, ppl=41.08
+2026-04-09 20:03:27 - INFO - parrotllm.training -   Val:   loss=3.6574, ppl=38.76
+2026-04-09 20:03:27 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:03:28 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p6574_epoch_0000_step_0002300.pt
+2026-04-09 20:03:29 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:03:32 - INFO - parrotllm.training - step   2310 | epoch 0 | loss 3.7336 | lr 2.49e-03 | grad 0.1620
+2026-04-09 20:03:35 - INFO - parrotllm.training - step   2320 | epoch 0 | loss 3.6906 | lr 2.46e-03 | grad 0.1584
+2026-04-09 20:03:38 - INFO - parrotllm.training - step   2330 | epoch 0 | loss 3.5583 | lr 2.43e-03 | grad 0.1468
+2026-04-09 20:03:41 - INFO - parrotllm.training - step   2340 | epoch 0 | loss 3.7180 | lr 2.40e-03 | grad 0.1823
+2026-04-09 20:03:44 - INFO - parrotllm.training - step   2350 | epoch 0 | loss 3.7389 | lr 2.37e-03 | grad 0.1359
+2026-04-09 20:03:47 - INFO - parrotllm.training - step   2360 | epoch 0 | loss 3.7068 | lr 2.34e-03 | grad 0.1608
+2026-04-09 20:03:50 - INFO - parrotllm.training - step   2370 | epoch 0 | loss 3.6860 | lr 2.31e-03 | grad 0.1505
+2026-04-09 20:03:53 - INFO - parrotllm.training - step   2380 | epoch 0 | loss 3.6225 | lr 2.28e-03 | grad 0.1429
+2026-04-09 20:03:56 - INFO - parrotllm.training - step   2390 | epoch 0 | loss 3.6969 | lr 2.25e-03 | grad 0.1640
+2026-04-09 20:03:59 - INFO - parrotllm.training - step   2400 | epoch 0 | loss 3.7678 | lr 2.22e-03 | grad 0.1342
+2026-04-09 20:03:59 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:03:59 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:04:04 - INFO - parrotllm.training -   Train: loss=3.7678, ppl=43.28
+2026-04-09 20:04:04 - INFO - parrotllm.training -   Val:   loss=3.6349, ppl=37.90
+2026-04-09 20:04:04 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:04:05 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p6349_epoch_0000_step_0002400.pt
+2026-04-09 20:04:06 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:04:09 - INFO - parrotllm.training - step   2410 | epoch 0 | loss 3.7203 | lr 2.19e-03 | grad 0.1528
+2026-04-09 20:04:12 - INFO - parrotllm.training - step   2420 | epoch 0 | loss 3.6639 | lr 2.16e-03 | grad 0.1403
+2026-04-09 20:04:15 - INFO - parrotllm.training - step   2430 | epoch 0 | loss 3.5508 | lr 2.13e-03 | grad 0.1437
+2026-04-09 20:04:18 - INFO - parrotllm.training - step   2440 | epoch 0 | loss 3.6180 | lr 2.10e-03 | grad 0.1430
+2026-04-09 20:04:21 - INFO - parrotllm.training - step   2450 | epoch 0 | loss 3.7059 | lr 2.07e-03 | grad 0.1310
+2026-04-09 20:04:24 - INFO - parrotllm.training - step   2460 | epoch 0 | loss 3.6396 | lr 2.04e-03 | grad 0.1581
+2026-04-09 20:04:27 - INFO - parrotllm.training - step   2470 | epoch 0 | loss 3.6486 | lr 2.01e-03 | grad 0.1315
+2026-04-09 20:04:30 - INFO - parrotllm.training - step   2480 | epoch 0 | loss 3.6033 | lr 1.98e-03 | grad 0.1561
+2026-04-09 20:04:33 - INFO - parrotllm.training - step   2490 | epoch 0 | loss 3.6343 | lr 1.95e-03 | grad 0.1392
+2026-04-09 20:04:36 - INFO - parrotllm.training - step   2500 | epoch 0 | loss 3.6688 | lr 1.92e-03 | grad 0.1552
+2026-04-09 20:04:36 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:04:36 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:04:40 - INFO - parrotllm.training -   Train: loss=3.6688, ppl=39.20
+2026-04-09 20:04:40 - INFO - parrotllm.training -   Val:   loss=3.6123, ppl=37.05
+2026-04-09 20:04:40 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:04:41 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p6123_epoch_0000_step_0002500.pt
+2026-04-09 20:04:42 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:04:43 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0002500.pt
+2026-04-09 20:04:47 - INFO - parrotllm.training - step   2510 | epoch 0 | loss 3.7239 | lr 1.89e-03 | grad 0.1364
+2026-04-09 20:04:50 - INFO - parrotllm.training - step   2520 | epoch 0 | loss 3.6800 | lr 1.86e-03 | grad 0.1355
+2026-04-09 20:04:53 - INFO - parrotllm.training - step   2530 | epoch 0 | loss 3.7082 | lr 1.83e-03 | grad 0.1309
+2026-04-09 20:04:56 - INFO - parrotllm.training - step   2540 | epoch 0 | loss 3.6512 | lr 1.80e-03 | grad 0.1308
+2026-04-09 20:04:59 - INFO - parrotllm.training - step   2550 | epoch 0 | loss 3.5392 | lr 1.77e-03 | grad 0.1426
+2026-04-09 20:05:02 - INFO - parrotllm.training - step   2560 | epoch 0 | loss 3.6647 | lr 1.74e-03 | grad 0.1390
+2026-04-09 20:05:06 - INFO - parrotllm.training - step   2570 | epoch 0 | loss 3.5630 | lr 1.71e-03 | grad 0.1394
+2026-04-09 20:05:09 - INFO - parrotllm.training - step   2580 | epoch 0 | loss 3.5162 | lr 1.68e-03 | grad 0.1409
+2026-04-09 20:05:12 - INFO - parrotllm.training - step   2590 | epoch 0 | loss 3.5925 | lr 1.65e-03 | grad 0.1300
+2026-04-09 20:05:15 - INFO - parrotllm.training - step   2600 | epoch 0 | loss 3.5902 | lr 1.62e-03 | grad 0.1330
+2026-04-09 20:05:15 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:05:15 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:05:19 - INFO - parrotllm.training -   Train: loss=3.5902, ppl=36.24
+2026-04-09 20:05:19 - INFO - parrotllm.training -   Val:   loss=3.5883, ppl=36.17
+2026-04-09 20:05:19 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:05:20 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p5883_epoch_0000_step_0002600.pt
+2026-04-09 20:05:21 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:05:24 - INFO - parrotllm.training - step   2610 | epoch 0 | loss 3.5127 | lr 1.59e-03 | grad 0.1256
+2026-04-09 20:05:27 - INFO - parrotllm.training - step   2620 | epoch 0 | loss 3.5998 | lr 1.56e-03 | grad 0.1246
+2026-04-09 20:05:30 - INFO - parrotllm.training - step   2630 | epoch 0 | loss 3.6302 | lr 1.53e-03 | grad 0.1386
+2026-04-09 20:05:33 - INFO - parrotllm.training - step   2640 | epoch 0 | loss 3.5715 | lr 1.50e-03 | grad 0.1244
+2026-04-09 20:05:36 - INFO - parrotllm.training - step   2650 | epoch 0 | loss 3.5813 | lr 1.47e-03 | grad 0.1272
+2026-04-09 20:05:39 - INFO - parrotllm.training - step   2660 | epoch 0 | loss 3.5388 | lr 1.44e-03 | grad 0.1316
+2026-04-09 20:05:42 - INFO - parrotllm.training - step   2670 | epoch 0 | loss 3.6599 | lr 1.41e-03 | grad 0.1381
+2026-04-09 20:05:45 - INFO - parrotllm.training - step   2680 | epoch 0 | loss 3.6654 | lr 1.38e-03 | grad 0.1162
+2026-04-09 20:05:48 - INFO - parrotllm.training - step   2690 | epoch 0 | loss 3.5500 | lr 1.35e-03 | grad 0.1230
+2026-04-09 20:05:51 - INFO - parrotllm.training - step   2700 | epoch 0 | loss 3.5983 | lr 1.32e-03 | grad 0.1323
+2026-04-09 20:05:51 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:05:51 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:05:56 - INFO - parrotllm.training -   Train: loss=3.5983, ppl=36.54
+2026-04-09 20:05:56 - INFO - parrotllm.training -   Val:   loss=3.5626, ppl=35.25
+2026-04-09 20:05:56 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:05:56 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p5626_epoch_0000_step_0002700.pt
+2026-04-09 20:05:57 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:06:00 - INFO - parrotllm.training - step   2710 | epoch 0 | loss 3.5847 | lr 1.29e-03 | grad 0.1429
+2026-04-09 20:06:03 - INFO - parrotllm.training - step   2720 | epoch 0 | loss 3.5221 | lr 1.26e-03 | grad 0.1347
+2026-04-09 20:06:06 - INFO - parrotllm.training - step   2730 | epoch 0 | loss 3.5486 | lr 1.23e-03 | grad 0.1298
+2026-04-09 20:06:09 - INFO - parrotllm.training - step   2740 | epoch 0 | loss 3.6082 | lr 1.20e-03 | grad 0.1197
+2026-04-09 20:06:12 - INFO - parrotllm.training - step   2750 | epoch 0 | loss 3.6150 | lr 1.17e-03 | grad 0.1362
+2026-04-09 20:06:13 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0002750.pt
+2026-04-09 20:06:17 - INFO - parrotllm.training - step   2760 | epoch 0 | loss 3.5811 | lr 1.14e-03 | grad 0.1395
+2026-04-09 20:06:20 - INFO - parrotllm.training - step   2770 | epoch 0 | loss 3.5431 | lr 1.11e-03 | grad 0.1306
+2026-04-09 20:06:23 - INFO - parrotllm.training - step   2780 | epoch 0 | loss 3.6263 | lr 1.08e-03 | grad 0.1222
+2026-04-09 20:06:26 - INFO - parrotllm.training - step   2790 | epoch 0 | loss 3.5704 | lr 1.05e-03 | grad 0.1321
+2026-04-09 20:06:29 - INFO - parrotllm.training - step   2800 | epoch 0 | loss 3.6078 | lr 1.02e-03 | grad 0.1325
+2026-04-09 20:06:29 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:06:29 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:06:34 - INFO - parrotllm.training -   Train: loss=3.6078, ppl=36.88
+2026-04-09 20:06:34 - INFO - parrotllm.training -   Val:   loss=3.5390, ppl=34.43
+2026-04-09 20:06:34 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:06:35 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p5390_epoch_0000_step_0002800.pt
+2026-04-09 20:06:36 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:06:39 - INFO - parrotllm.training - step   2810 | epoch 0 | loss 3.5558 | lr 9.90e-04 | grad 0.1296
+2026-04-09 20:06:42 - INFO - parrotllm.training - step   2820 | epoch 0 | loss 3.5991 | lr 9.60e-04 | grad 0.1256
+2026-04-09 20:06:45 - INFO - parrotllm.training - step   2830 | epoch 0 | loss 3.5413 | lr 9.30e-04 | grad 0.1221
+2026-04-09 20:06:48 - INFO - parrotllm.training - step   2840 | epoch 0 | loss 3.6000 | lr 9.00e-04 | grad 0.1217
+2026-04-09 20:06:51 - INFO - parrotllm.training - step   2850 | epoch 0 | loss 3.5173 | lr 8.69e-04 | grad 0.1250
+2026-04-09 20:06:54 - INFO - parrotllm.training - step   2860 | epoch 0 | loss 3.5975 | lr 8.39e-04 | grad 0.1222
+2026-04-09 20:06:57 - INFO - parrotllm.training - step   2870 | epoch 0 | loss 3.5660 | lr 8.09e-04 | grad 0.1181
+2026-04-09 20:07:00 - INFO - parrotllm.training - step   2880 | epoch 0 | loss 3.5448 | lr 7.79e-04 | grad 0.1156
+2026-04-09 20:07:03 - INFO - parrotllm.training - step   2890 | epoch 0 | loss 3.6903 | lr 7.49e-04 | grad 0.1265
+2026-04-09 20:07:06 - INFO - parrotllm.training - step   2900 | epoch 0 | loss 3.5977 | lr 7.19e-04 | grad 0.1205
+2026-04-09 20:07:06 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:07:06 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:07:10 - INFO - parrotllm.training -   Train: loss=3.5977, ppl=36.52
+2026-04-09 20:07:10 - INFO - parrotllm.training -   Val:   loss=3.5182, ppl=33.72
+2026-04-09 20:07:10 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:07:11 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p5182_epoch_0000_step_0002900.pt
+2026-04-09 20:07:12 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:07:15 - INFO - parrotllm.training - step   2910 | epoch 0 | loss 3.5500 | lr 6.89e-04 | grad 0.1208
+2026-04-09 20:07:18 - INFO - parrotllm.training - step   2920 | epoch 0 | loss 3.5258 | lr 6.59e-04 | grad 0.1128
+2026-04-09 20:07:21 - INFO - parrotllm.training - step   2930 | epoch 0 | loss 3.5937 | lr 6.29e-04 | grad 0.1165
+2026-04-09 20:07:24 - INFO - parrotllm.training - step   2940 | epoch 0 | loss 3.4018 | lr 5.99e-04 | grad 0.1168
+2026-04-09 20:07:28 - INFO - parrotllm.training - step   2950 | epoch 0 | loss 3.4564 | lr 5.69e-04 | grad 0.1157
+2026-04-09 20:07:31 - INFO - parrotllm.training - step   2960 | epoch 0 | loss 3.5687 | lr 5.39e-04 | grad 0.1149
+2026-04-09 20:07:34 - INFO - parrotllm.training - step   2970 | epoch 0 | loss 3.5264 | lr 5.09e-04 | grad 0.1211
+2026-04-09 20:07:37 - INFO - parrotllm.training - step   2980 | epoch 0 | loss 3.4588 | lr 4.79e-04 | grad 0.1173
+2026-04-09 20:07:40 - INFO - parrotllm.training - step   2990 | epoch 0 | loss 3.5801 | lr 4.49e-04 | grad 0.1131
+2026-04-09 20:07:43 - INFO - parrotllm.training - step   3000 | epoch 0 | loss 3.5708 | lr 4.19e-04 | grad 0.1195
+2026-04-09 20:07:43 - INFO - parrotllm.training - Starting evaluation...
+2026-04-09 20:07:43 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:07:47 - INFO - parrotllm.training -   Train: loss=3.5708, ppl=35.54
+2026-04-09 20:07:47 - INFO - parrotllm.training -   Val:   loss=3.4954, ppl=32.96
+2026-04-09 20:07:47 - INFO - parrotllm.training -   ** New best validation loss! **
+2026-04-09 20:07:48 - INFO - parrotllm.training -   Saved best checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/best_loss_3p4954_epoch_0000_step_0003000.pt
+2026-04-09 20:07:49 - INFO - parrotllm.training - ------------------------------------------------------------
+2026-04-09 20:07:50 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0003000.pt
+2026-04-09 20:07:54 - INFO - parrotllm.training - step   3010 | epoch 0 | loss 3.5913 | lr 3.89e-04 | grad 0.1064
+2026-04-09 20:07:57 - INFO - parrotllm.training - step   3020 | epoch 0 | loss 3.4722 | lr 3.59e-04 | grad 0.1119
+2026-04-09 20:08:00 - INFO - parrotllm.training - step   3030 | epoch 0 | loss 3.5586 | lr 3.29e-04 | grad 0.1068
+2026-04-09 20:08:03 - INFO - parrotllm.training - step   3040 | epoch 0 | loss 3.5189 | lr 2.99e-04 | grad 0.1113
+2026-04-09 20:08:06 - INFO - parrotllm.training - step   3050 | epoch 0 | loss 3.5322 | lr 2.69e-04 | grad 0.1048
+2026-04-09 20:08:08 - INFO - parrotllm.training - Saved final checkpoint: /home/user/code/ParrotLLM/runs/big_run/exp_c/run_20260409_194841/checkpoints/last_epoch_0000_step_0003052.pt
+2026-04-09 20:08:08 - INFO - parrotllm.training - 
+Loss Curve:
+  Loss
+  ^
+   10.91 |*                                                 
+         |*                                                 
+         |*                                                 
+         |*                                                 
+         |**                                                
+         | ****                                             
+         |   ******                                         
+    3.40 |       *******************************************
+         +--------------------------------------------------> step
+         1                                             3052
+2026-04-09 20:08:08 - INFO - parrotllm.training - 
+============================================================
+TRAINING COMPLETE
+============================================================
+  Epochs: 1
+  Total steps: 3052
+  Total time: 0.32 hours
+  Best validation loss: 3.4954
+  Run directory: runs/big_run/exp_c/run_20260409_194841
+============================================================
+2026-04-09 20:08:08 - INFO - parrotllm.training - Uploading completed run directory to Hugging Face: repo=ParrotLabs/Preprocessed

runs/big_run/exp_c/run_20260409_194841/training_plots.pdf

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4547c027e95e0072a9db84539f088d77b30f0866ade5e3c8de1aefb8851806bb
+size 112296