Hugging Face's logo

Datasets:
ParrotLabs
/
Preprocessed

Dataset card Files
xet
 Community
thaferbe commited on
Commit
ee91277
·
verified ·
1 Parent(s): 9381c18

Upload training run run_20260408_102350

Browse files

ParrotLLM end-of-training sync for local run runs/fast_learn/run_20260408_102350.

Files changed (14) hide show
  1. runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p4871_epoch_0005_step_0000290.pt +3 -0
  2. runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5027_epoch_0004_step_0000280.pt +3 -0
  3. runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5059_epoch_0004_step_0000270.pt +3 -0
  4. runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5104_epoch_0005_step_0000300.pt +3 -0
  5. runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5239_epoch_0004_step_0000240.pt +3 -0
  6. runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0004_step_0000260.pt +3 -0
  7. runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0004_step_0000270.pt +3 -0
  8. runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0004_step_0000280.pt +3 -0
  9. runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0005_step_0000290.pt +3 -0
  10. runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0005_step_0000300.pt +3 -0
  11. runs/fast_learn/run_20260408_102350/config.json +68 -0
  12. runs/fast_learn/run_20260408_102350/metrics.jsonl +0 -0
  13. runs/fast_learn/run_20260408_102350/train.log +418 -0
  14. runs/fast_learn/run_20260408_102350/training_plots.pdf +0 -0
runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p4871_epoch_0005_step_0000290.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f7e3f5efa1636de6a02688c5c5d023c8cc6dd6c91318fce65ed3a4068b1757b7
3
+ size 192265028
runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5027_epoch_0004_step_0000280.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:7cb309240a642a59af830dca22d3b71c048b20b726a3a4d689a0e671eb2633e4
3
+ size 192265028
runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5059_epoch_0004_step_0000270.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:ef7ea9b6c93a1598bd9a7a35e89fcc80d2f84e6da77cf67077acae5e63607330
3
+ size 192265028
runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5104_epoch_0005_step_0000300.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:05ace2325f66db9fb4db76394487eb4a2b7e3a63091750aff559c0ec7e7d9aba
3
+ size 192265028
runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5239_epoch_0004_step_0000240.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:8d37c55f6504221bb3cd54cf55c14f8defac22cd585c55b90416692a972870c8
3
+ size 192265028
runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0004_step_0000260.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3c47ec975e49098dc9f80085be2f2c585802c809aa60db20ed35e5adbce92846
3
+ size 192262192
runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0004_step_0000270.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f5fb70c04be6173388015edbc63e91d1bbbdc5e371fbc5a8f8b755b02266eb95
3
+ size 192262192
runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0004_step_0000280.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:e30539e05e9c91ae0eb013be7f54803d6b2ba1c2e20ea138dc1486acad5928c7
3
+ size 192262192
runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0005_step_0000290.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:82e14e83fb4586a2c5b24c68f18dde1a82604ca4942a8635126eb0c9a962e9e7
3
+ size 192262192
runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0005_step_0000300.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6ae233b6c8aa7949e0339fc804f74fb2202fd633ff3a705e2409e1bdf2fa3ec0
3
+ size 192262192
runs/fast_learn/run_20260408_102350/config.json ADDED
@@ -0,0 +1,68 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "preprocess": null,
3
+ "model": {
4
+ "vocab_size": 50258,
5
+ "pad_token_id": 50257,
6
+ "bos_token_id": 50256,
7
+ "eos_token_id": 50256,
8
+ "d_model": 256,
9
+ "n_layers": 4,
10
+ "n_heads": 8,
11
+ "d_ff": 682,
12
+ "context_length": 256,
13
+ "bias": false,
14
+ "dropout": 0.0,
15
+ "rope_theta": 10000.0,
16
+ "gradient_checkpointing": false
17
+ },
18
+ "training": {
19
+ "device": "auto",
20
+ "train_bin": "data/SmokeTest/train.bin",
21
+ "val_bin": "data/SmokeTest/val.bin",
22
+ "num_workers": 0,
23
+ "pin_memory": true,
24
+ "batch_size": 64,
25
+ "gradient_accumulation_steps": 1,
26
+ "learning_rate": 0.001,
27
+ "min_lr": 0.0001,
28
+ "weight_decay": 0.1,
29
+ "beta1": 0.9,
30
+ "beta2": 0.95,
31
+ "grad_clip": 1.0,
32
+ "warmup_steps": 20,
33
+ "max_steps": 300,
34
+ "lr_schedule": "wsd",
35
+ "lr_decay_ratio": 0.2,
36
+ "z_loss_coeff": 0.0001,
37
+ "save_every": 10,
38
+ "eval_every": 10,
39
+ "checkpoint_dir": "checkpoints",
40
+ "keep_last_checkpoints": 5,
41
+ "keep_best_checkpoints": 5,
42
+ "early_stopping_patience": 3,
43
+ "early_stopping_min_delta": 0.0,
44
+ "runs_dir": "runs/fast_learn",
45
+ "log_every": 10,
46
+ "hf_upload": {
47
+ "repo_id": "ParrotLabs/Preprocessed",
48
+ "repo_type": "dataset",
49
+ "path_in_repo": "",
50
+ "private": null
51
+ },
52
+ "compile": false
53
+ },
54
+ "tune": null,
55
+ "logging": {
56
+ "console_level": "INFO",
57
+ "file_level": "DEBUG",
58
+ "components": {
59
+ "data_preprocessing": "INFO",
60
+ "model_initialization": "INFO",
61
+ "training": "INFO"
62
+ },
63
+ "profiler": null
64
+ },
65
+ "eval": null,
66
+ "inference": null,
67
+ "chat": null
68
+ }
runs/fast_learn/run_20260408_102350/metrics.jsonl ADDED
The diff for this file is too large to render. See raw diff
 
runs/fast_learn/run_20260408_102350/train.log ADDED
@@ -0,0 +1,418 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-04-08 10:23:50 - INFO - parrotllm - Logging initialised -> runs/fast_learn/run_20260408_102350/train.log
2
+ 2026-04-08 10:23:50 - INFO - parrotllm.training - device=cuda | rank=0 | world_size=1 | distributed=no
3
+ 2026-04-08 10:23:50 - INFO - parrotllm.training - checkpoint directory=/home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints | keep_last=5 | keep_best=5
4
+ 2026-04-08 10:23:51 - INFO - parrotllm.training - torch.compile disabled via config
5
+ 2026-04-08 10:23:51 - INFO - parrotllm.training -
6
+ ============================================================
7
+ MODEL ARCHITECTURE SUMMARY
8
+ ============================================================
9
+
10
+ Configuration:
11
+ Vocab size: 50258
12
+ Block size (context): 256
13
+ Layers: 4
14
+ Heads: 8
15
+ Embedding dim: 256
16
+ FFN hidden dim: 682
17
+ Dropout: 0.0
18
+ Bias: False
19
+
20
+ Parameters (unique, weight-tied layers counted once):
21
+ Total: 16,014,336 (16.01M)
22
+ Trainable: 16,014,336 (16.01M)
23
+ Non-trainable: 0
24
+ Non-embedding: 16,014,336 (16.01M)
25
+ Position embeddings: 0 (RoPE: 0 learned params)
26
+ Size (MB): 64.06
27
+
28
+ Layer-wise breakdown (torchinfo):
29
+ Note: torchinfo double-counts weight-tied layers (tok_emb/lm_head).
30
+ ============================================================================================================================================
31
+ Layer (type:depth-idx) Input Shape Output Shape Param # Trainable
32
+ ============================================================================================================================================
33
+ ParrotLLM [64, 256] [64, 256, 50258] -- True
34
+ ├─Embedding: 1-1 [64, 256] [64, 256, 256] 12,866,048 True
35
+ ├─Dropout: 1-2 [64, 256, 256] [64, 256, 256] -- --
36
+ ├─ModuleList: 1-3 -- -- -- True
37
+ │ └─TransformerBlock: 2-1 [64, 256, 256] [64, 256, 256] -- True
38
+ │ │ └─RMSNorm: 3-1 [64, 256, 256] [64, 256, 256] 256 True
39
+ │ │ └─MultiHeadAttention: 3-2 [64, 256, 256] [64, 256, 256] 262,208 True
40
+ │ │ └─RMSNorm: 3-3 [64, 256, 256] [64, 256, 256] 256 True
41
+ │ │ └─RMSNorm: 3-4 [64, 256, 256] [64, 256, 256] 256 True
42
+ │ │ └─SwiGLUMLP: 3-5 [64, 256, 256] [64, 256, 256] 523,776 True
43
+ │ │ └─RMSNorm: 3-6 [64, 256, 256] [64, 256, 256] 256 True
44
+ │ └─TransformerBlock: 2-2 [64, 256, 256] [64, 256, 256] -- True
45
+ │ │ └─RMSNorm: 3-7 [64, 256, 256] [64, 256, 256] 256 True
46
+ │ │ └─MultiHeadAttention: 3-8 [64, 256, 256] [64, 256, 256] 262,208 True
47
+ │ │ └─RMSNorm: 3-9 [64, 256, 256] [64, 256, 256] 256 True
48
+ │ │ └─RMSNorm: 3-10 [64, 256, 256] [64, 256, 256] 256 True
49
+ │ │ └─SwiGLUMLP: 3-11 [64, 256, 256] [64, 256, 256] 523,776 True
50
+ │ │ └─RMSNorm: 3-12 [64, 256, 256] [64, 256, 256] 256 True
51
+ │ └─TransformerBlock: 2-3 [64, 256, 256] [64, 256, 256] -- True
52
+ │ │ └─RMSNorm: 3-13 [64, 256, 256] [64, 256, 256] 256 True
53
+ │ │ └─MultiHeadAttention: 3-14 [64, 256, 256] [64, 256, 256] 262,208 True
54
+ │ │ └─RMSNorm: 3-15 [64, 256, 256] [64, 256, 256] 256 True
55
+ │ │ └─RMSNorm: 3-16 [64, 256, 256] [64, 256, 256] 256 True
56
+ │ │ └─SwiGLUMLP: 3-17 [64, 256, 256] [64, 256, 256] 523,776 True
57
+ │ │ └─RMSNorm: 3-18 [64, 256, 256] [64, 256, 256] 256 True
58
+ │ └─TransformerBlock: 2-4 [64, 256, 256] [64, 256, 256] -- True
59
+ │ │ └─RMSNorm: 3-19 [64, 256, 256] [64, 256, 256] 256 True
60
+ │ │ └─MultiHeadAttention: 3-20 [64, 256, 256] [64, 256, 256] 262,208 True
61
+ │ │ └─RMSNorm: 3-21 [64, 256, 256] [64, 256, 256] 256 True
62
+ │ │ └─RMSNorm: 3-22 [64, 256, 256] [64, 256, 256] 256 True
63
+ │ │ └─SwiGLUMLP: 3-23 [64, 256, 256] [64, 256, 256] 523,776 True
64
+ │ │ └─RMSNorm: 3-24 [64, 256, 256] [64, 256, 256] 256 True
65
+ ├─RMSNorm: 1-4 [64, 256, 256] [64, 256, 256] 256 True
66
+ ├─Linear: 1-5 [64, 256, 256] [64, 256, 50258] 12,866,048 True
67
+ ============================================================================================================================================
68
+ Total params: 28,880,384
69
+ Trainable params: 28,880,384
70
+ Non-trainable params: 0
71
+ Total mult-adds (Units.GIGABYTES): 1.85
72
+ ============================================================================================================================================
73
+ Input size (MB): 0.13
74
+ Forward/backward pass size (MB): 8846.05
75
+ Params size (MB): 115.52
76
+ Estimated Total Size (MB): 8961.70
77
+ ============================================================================================================================================
78
+ ============================================================
79
+ 2026-04-08 10:23:51 - INFO - parrotllm.training - Starting evaluation...
80
+ 2026-04-08 10:23:51 - INFO - parrotllm.training - Initial val: loss=10.8621, ppl=52163.59
81
+ 2026-04-08 10:23:51 - INFO - parrotllm.training -
82
+ ============================================================
83
+ Starting training...
84
+ Steps per epoch (approx): 58
85
+ Max steps: 300
86
+ ============================================================
87
+ 2026-04-08 10:23:51 - INFO - parrotllm.training - LR schedule=wsd | decay_ratio=0.2 | z_loss_coeff=0.0001
88
+ 2026-04-08 10:23:51 - INFO - parrotllm.training - Early stopping enabled | patience=3 evals | min_delta=0.000000
89
+ 2026-04-08 10:23:51 - INFO - parrotllm.training - step 1 | epoch 0 | loss 10.8775 | lr 5.00e-05 | grad 2.4289
90
+ 2026-04-08 10:23:53 - INFO - parrotllm.training - step 10 | epoch 0 | loss 10.0071 | lr 5.00e-04 | grad 1.3299
91
+ 2026-04-08 10:23:53 - INFO - parrotllm.training - Starting evaluation...
92
+ 2026-04-08 10:23:53 - INFO - parrotllm.training - ------------------------------------------------------------
93
+ 2026-04-08 10:23:53 - INFO - parrotllm.training - Train: loss=10.0071, ppl=22183.73
94
+ 2026-04-08 10:23:53 - INFO - parrotllm.training - Val: loss=9.8934, ppl=19799.94
95
+ 2026-04-08 10:23:53 - INFO - parrotllm.training - ** New best validation loss! **
96
+ 2026-04-08 10:23:53 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_9p8934_epoch_0000_step_0000010.pt
97
+ 2026-04-08 10:23:54 - INFO - parrotllm.training - ------------------------------------------------------------
98
+ 2026-04-08 10:23:55 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0000_step_0000010.pt
99
+ 2026-04-08 10:23:57 - INFO - parrotllm.training - step 20 | epoch 0 | loss 8.4109 | lr 1.00e-03 | grad 1.2339
100
+ 2026-04-08 10:23:57 - INFO - parrotllm.training - Starting evaluation...
101
+ 2026-04-08 10:23:57 - INFO - parrotllm.training - ------------------------------------------------------------
102
+ 2026-04-08 10:23:57 - INFO - parrotllm.training - Train: loss=8.4109, ppl=4495.78
103
+ 2026-04-08 10:23:57 - INFO - parrotllm.training - Val: loss=8.3807, ppl=4361.91
104
+ 2026-04-08 10:23:57 - INFO - parrotllm.training - ** New best validation loss! **
105
+ 2026-04-08 10:23:57 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_8p3807_epoch_0000_step_0000020.pt
106
+ 2026-04-08 10:23:58 - INFO - parrotllm.training - ------------------------------------------------------------
107
+ 2026-04-08 10:23:58 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0000_step_0000020.pt
108
+ 2026-04-08 10:24:00 - INFO - parrotllm.training - step 30 | epoch 0 | loss 7.5825 | lr 1.00e-03 | grad 0.4779
109
+ 2026-04-08 10:24:00 - INFO - parrotllm.training - Starting evaluation...
110
+ 2026-04-08 10:24:00 - INFO - parrotllm.training - ------------------------------------------------------------
111
+ 2026-04-08 10:24:01 - INFO - parrotllm.training - Train: loss=7.5825, ppl=1963.53
112
+ 2026-04-08 10:24:01 - INFO - parrotllm.training - Val: loss=7.7298, ppl=2275.08
113
+ 2026-04-08 10:24:01 - INFO - parrotllm.training - ** New best validation loss! **
114
+ 2026-04-08 10:24:01 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_7p7298_epoch_0000_step_0000030.pt
115
+ 2026-04-08 10:24:02 - INFO - parrotllm.training - ------------------------------------------------------------
116
+ 2026-04-08 10:24:02 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0000_step_0000030.pt
117
+ 2026-04-08 10:24:04 - INFO - parrotllm.training - step 40 | epoch 0 | loss 7.2955 | lr 1.00e-03 | grad 0.4186
118
+ 2026-04-08 10:24:04 - INFO - parrotllm.training - Starting evaluation...
119
+ 2026-04-08 10:24:04 - INFO - parrotllm.training - ------------------------------------------------------------
120
+ 2026-04-08 10:24:04 - INFO - parrotllm.training - Train: loss=7.2955, ppl=1473.72
121
+ 2026-04-08 10:24:04 - INFO - parrotllm.training - Val: loss=7.5845, ppl=1967.48
122
+ 2026-04-08 10:24:04 - INFO - parrotllm.training - ** New best validation loss! **
123
+ 2026-04-08 10:24:05 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_7p5845_epoch_0000_step_0000040.pt
124
+ 2026-04-08 10:24:06 - INFO - parrotllm.training - ------------------------------------------------------------
125
+ 2026-04-08 10:24:06 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0000_step_0000040.pt
126
+ 2026-04-08 10:24:08 - INFO - parrotllm.training - step 50 | epoch 0 | loss 7.1519 | lr 1.00e-03 | grad 0.8343
127
+ 2026-04-08 10:24:08 - INFO - parrotllm.training - Starting evaluation...
128
+ 2026-04-08 10:24:08 - INFO - parrotllm.training - ------------------------------------------------------------
129
+ 2026-04-08 10:24:09 - INFO - parrotllm.training - Train: loss=7.1519, ppl=1276.52
130
+ 2026-04-08 10:24:09 - INFO - parrotllm.training - Val: loss=7.4119, ppl=1655.60
131
+ 2026-04-08 10:24:09 - INFO - parrotllm.training - ** New best validation loss! **
132
+ 2026-04-08 10:24:09 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_7p4119_epoch_0000_step_0000050.pt
133
+ 2026-04-08 10:24:10 - INFO - parrotllm.training - ------------------------------------------------------------
134
+ 2026-04-08 10:24:10 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0000_step_0000050.pt
135
+ 2026-04-08 10:24:12 - INFO - parrotllm.training - Starting evaluation...
136
+ 2026-04-08 10:24:12 - INFO - parrotllm.training - ------------------------------------------------------------
137
+ 2026-04-08 10:24:12 - INFO - parrotllm.training - Epoch 1 complete:
138
+ 2026-04-08 10:24:12 - INFO - parrotllm.training - Train: loss=7.0841, ppl=1192.81
139
+ 2026-04-08 10:24:12 - INFO - parrotllm.training - Val: loss=7.2928, ppl=1469.74
140
+ 2026-04-08 10:24:12 - INFO - parrotllm.training - ** New best validation loss! **
141
+ 2026-04-08 10:24:13 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_7p2928_epoch_0001_step_0000058.pt
142
+ 2026-04-08 10:24:13 - INFO - parrotllm.training - ------------------------------------------------------------
143
+ 2026-04-08 10:24:14 - INFO - parrotllm.training - step 60 | epoch 1 | loss 7.0422 | lr 1.00e-03 | grad 0.3758
144
+ 2026-04-08 10:24:14 - INFO - parrotllm.training - Starting evaluation...
145
+ 2026-04-08 10:24:14 - INFO - parrotllm.training - ------------------------------------------------------------
146
+ 2026-04-08 10:24:14 - INFO - parrotllm.training - Train: loss=7.0422, ppl=1143.94
147
+ 2026-04-08 10:24:14 - INFO - parrotllm.training - Val: loss=7.2751, ppl=1443.95
148
+ 2026-04-08 10:24:14 - INFO - parrotllm.training - ** New best validation loss! **
149
+ 2026-04-08 10:24:14 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_7p2751_epoch_0001_step_0000060.pt
150
+ 2026-04-08 10:24:15 - INFO - parrotllm.training - ------------------------------------------------------------
151
+ 2026-04-08 10:24:15 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0001_step_0000060.pt
152
+ 2026-04-08 10:24:17 - INFO - parrotllm.training - step 70 | epoch 1 | loss 6.7969 | lr 1.00e-03 | grad 0.3625
153
+ 2026-04-08 10:24:17 - INFO - parrotllm.training - Starting evaluation...
154
+ 2026-04-08 10:24:17 - INFO - parrotllm.training - ------------------------------------------------------------
155
+ 2026-04-08 10:24:18 - INFO - parrotllm.training - Train: loss=6.7969, ppl=895.10
156
+ 2026-04-08 10:24:18 - INFO - parrotllm.training - Val: loss=7.1586, ppl=1285.06
157
+ 2026-04-08 10:24:18 - INFO - parrotllm.training - ** New best validation loss! **
158
+ 2026-04-08 10:24:18 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_7p1586_epoch_0001_step_0000070.pt
159
+ 2026-04-08 10:24:19 - INFO - parrotllm.training - ------------------------------------------------------------
160
+ 2026-04-08 10:24:19 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0001_step_0000070.pt
161
+ 2026-04-08 10:24:21 - INFO - parrotllm.training - step 80 | epoch 1 | loss 6.7478 | lr 1.00e-03 | grad 0.4638
162
+ 2026-04-08 10:24:21 - INFO - parrotllm.training - Starting evaluation...
163
+ 2026-04-08 10:24:21 - INFO - parrotllm.training - ------------------------------------------------------------
164
+ 2026-04-08 10:24:21 - INFO - parrotllm.training - Train: loss=6.7478, ppl=852.20
165
+ 2026-04-08 10:24:21 - INFO - parrotllm.training - Val: loss=7.0772, ppl=1184.70
166
+ 2026-04-08 10:24:21 - INFO - parrotllm.training - ** New best validation loss! **
167
+ 2026-04-08 10:24:22 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_7p0772_epoch_0001_step_0000080.pt
168
+ 2026-04-08 10:24:23 - INFO - parrotllm.training - ------------------------------------------------------------
169
+ 2026-04-08 10:24:24 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0001_step_0000080.pt
170
+ 2026-04-08 10:24:26 - INFO - parrotllm.training - step 90 | epoch 1 | loss 6.6866 | lr 1.00e-03 | grad 0.3555
171
+ 2026-04-08 10:24:26 - INFO - parrotllm.training - Starting evaluation...
172
+ 2026-04-08 10:24:26 - INFO - parrotllm.training - ------------------------------------------------------------
173
+ 2026-04-08 10:24:26 - INFO - parrotllm.training - Train: loss=6.6866, ppl=801.63
174
+ 2026-04-08 10:24:26 - INFO - parrotllm.training - Val: loss=6.9723, ppl=1066.66
175
+ 2026-04-08 10:24:26 - INFO - parrotllm.training - ** New best validation loss! **
176
+ 2026-04-08 10:24:27 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p9723_epoch_0001_step_0000090.pt
177
+ 2026-04-08 10:24:27 - INFO - parrotllm.training - ------------------------------------------------------------
178
+ 2026-04-08 10:24:28 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0001_step_0000090.pt
179
+ 2026-04-08 10:24:31 - INFO - parrotllm.training - step 100 | epoch 1 | loss 6.5857 | lr 1.00e-03 | grad 0.4117
180
+ 2026-04-08 10:24:31 - INFO - parrotllm.training - Starting evaluation...
181
+ 2026-04-08 10:24:31 - INFO - parrotllm.training - ------------------------------------------------------------
182
+ 2026-04-08 10:24:31 - INFO - parrotllm.training - Train: loss=6.5857, ppl=724.68
183
+ 2026-04-08 10:24:31 - INFO - parrotllm.training - Val: loss=6.9045, ppl=996.71
184
+ 2026-04-08 10:24:31 - INFO - parrotllm.training - ** New best validation loss! **
185
+ 2026-04-08 10:24:32 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p9045_epoch_0001_step_0000100.pt
186
+ 2026-04-08 10:24:32 - INFO - parrotllm.training - ------------------------------------------------------------
187
+ 2026-04-08 10:24:33 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0001_step_0000100.pt
188
+ 2026-04-08 10:24:35 - INFO - parrotllm.training - step 110 | epoch 1 | loss 6.4655 | lr 1.00e-03 | grad 0.4674
189
+ 2026-04-08 10:24:35 - INFO - parrotllm.training - Starting evaluation...
190
+ 2026-04-08 10:24:35 - INFO - parrotllm.training - ------------------------------------------------------------
191
+ 2026-04-08 10:24:35 - INFO - parrotllm.training - Train: loss=6.4655, ppl=642.61
192
+ 2026-04-08 10:24:35 - INFO - parrotllm.training - Val: loss=6.8364, ppl=931.09
193
+ 2026-04-08 10:24:35 - INFO - parrotllm.training - ** New best validation loss! **
194
+ 2026-04-08 10:24:36 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p8364_epoch_0001_step_0000110.pt
195
+ 2026-04-08 10:24:37 - INFO - parrotllm.training - ------------------------------------------------------------
196
+ 2026-04-08 10:24:38 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0001_step_0000110.pt
197
+ 2026-04-08 10:24:39 - INFO - parrotllm.training - Starting evaluation...
198
+ 2026-04-08 10:24:39 - INFO - parrotllm.training - ------------------------------------------------------------
199
+ 2026-04-08 10:24:39 - INFO - parrotllm.training - Epoch 2 complete:
200
+ 2026-04-08 10:24:39 - INFO - parrotllm.training - Train: loss=6.5477, ppl=697.67
201
+ 2026-04-08 10:24:39 - INFO - parrotllm.training - Val: loss=6.8036, ppl=901.13
202
+ 2026-04-08 10:24:39 - INFO - parrotllm.training - ** New best validation loss! **
203
+ 2026-04-08 10:24:40 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p8036_epoch_0002_step_0000116.pt
204
+ 2026-04-08 10:24:41 - INFO - parrotllm.training - ------------------------------------------------------------
205
+ 2026-04-08 10:24:42 - INFO - parrotllm.training - step 120 | epoch 2 | loss 6.2526 | lr 1.00e-03 | grad 0.4435
206
+ 2026-04-08 10:24:42 - INFO - parrotllm.training - Starting evaluation...
207
+ 2026-04-08 10:24:42 - INFO - parrotllm.training - ------------------------------------------------------------
208
+ 2026-04-08 10:24:42 - INFO - parrotllm.training - Train: loss=6.2526, ppl=519.35
209
+ 2026-04-08 10:24:42 - INFO - parrotllm.training - Val: loss=6.7931, ppl=891.67
210
+ 2026-04-08 10:24:42 - INFO - parrotllm.training - ** New best validation loss! **
211
+ 2026-04-08 10:24:42 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p7931_epoch_0002_step_0000120.pt
212
+ 2026-04-08 10:24:43 - INFO - parrotllm.training - ------------------------------------------------------------
213
+ 2026-04-08 10:24:45 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0002_step_0000120.pt
214
+ 2026-04-08 10:24:47 - INFO - parrotllm.training - step 130 | epoch 2 | loss 6.2556 | lr 1.00e-03 | grad 0.4452
215
+ 2026-04-08 10:24:47 - INFO - parrotllm.training - Starting evaluation...
216
+ 2026-04-08 10:24:47 - INFO - parrotllm.training - ------------------------------------------------------------
217
+ 2026-04-08 10:24:47 - INFO - parrotllm.training - Train: loss=6.2556, ppl=520.94
218
+ 2026-04-08 10:24:47 - INFO - parrotllm.training - Val: loss=6.7678, ppl=869.40
219
+ 2026-04-08 10:24:47 - INFO - parrotllm.training - ** New best validation loss! **
220
+ 2026-04-08 10:24:48 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p7678_epoch_0002_step_0000130.pt
221
+ 2026-04-08 10:24:48 - INFO - parrotllm.training - ------------------------------------------------------------
222
+ 2026-04-08 10:24:49 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0002_step_0000130.pt
223
+ 2026-04-08 10:24:51 - INFO - parrotllm.training - step 140 | epoch 2 | loss 6.3007 | lr 1.00e-03 | grad 0.4902
224
+ 2026-04-08 10:24:51 - INFO - parrotllm.training - Starting evaluation...
225
+ 2026-04-08 10:24:51 - INFO - parrotllm.training - ------------------------------------------------------------
226
+ 2026-04-08 10:24:51 - INFO - parrotllm.training - Train: loss=6.3007, ppl=544.97
227
+ 2026-04-08 10:24:51 - INFO - parrotllm.training - Val: loss=6.7293, ppl=836.56
228
+ 2026-04-08 10:24:51 - INFO - parrotllm.training - ** New best validation loss! **
229
+ 2026-04-08 10:24:51 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p7293_epoch_0002_step_0000140.pt
230
+ 2026-04-08 10:24:52 - INFO - parrotllm.training - ------------------------------------------------------------
231
+ 2026-04-08 10:24:53 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0002_step_0000140.pt
232
+ 2026-04-08 10:24:55 - INFO - parrotllm.training - step 150 | epoch 2 | loss 6.1964 | lr 1.00e-03 | grad 0.5096
233
+ 2026-04-08 10:24:55 - INFO - parrotllm.training - Starting evaluation...
234
+ 2026-04-08 10:24:55 - INFO - parrotllm.training - ------------------------------------------------------------
235
+ 2026-04-08 10:24:55 - INFO - parrotllm.training - Train: loss=6.1964, ppl=490.96
236
+ 2026-04-08 10:24:55 - INFO - parrotllm.training - Val: loss=6.6882, ppl=802.88
237
+ 2026-04-08 10:24:55 - INFO - parrotllm.training - ** New best validation loss! **
238
+ 2026-04-08 10:24:55 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p6882_epoch_0002_step_0000150.pt
239
+ 2026-04-08 10:24:56 - INFO - parrotllm.training - ------------------------------------------------------------
240
+ 2026-04-08 10:24:57 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0002_step_0000150.pt
241
+ 2026-04-08 10:24:59 - INFO - parrotllm.training - step 160 | epoch 2 | loss 6.1406 | lr 1.00e-03 | grad 0.5823
242
+ 2026-04-08 10:24:59 - INFO - parrotllm.training - Starting evaluation...
243
+ 2026-04-08 10:24:59 - INFO - parrotllm.training - ------------------------------------------------------------
244
+ 2026-04-08 10:24:59 - INFO - parrotllm.training - Train: loss=6.1406, ppl=464.32
245
+ 2026-04-08 10:24:59 - INFO - parrotllm.training - Val: loss=6.6567, ppl=777.97
246
+ 2026-04-08 10:24:59 - INFO - parrotllm.training - ** New best validation loss! **
247
+ 2026-04-08 10:25:00 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p6567_epoch_0002_step_0000160.pt
248
+ 2026-04-08 10:25:01 - INFO - parrotllm.training - ------------------------------------------------------------
249
+ 2026-04-08 10:25:01 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0002_step_0000160.pt
250
+ 2026-04-08 10:25:04 - INFO - parrotllm.training - step 170 | epoch 2 | loss 6.1889 | lr 1.00e-03 | grad 0.5073
251
+ 2026-04-08 10:25:04 - INFO - parrotllm.training - Starting evaluation...
252
+ 2026-04-08 10:25:04 - INFO - parrotllm.training - ------------------------------------------------------------
253
+ 2026-04-08 10:25:04 - INFO - parrotllm.training - Train: loss=6.1889, ppl=487.33
254
+ 2026-04-08 10:25:04 - INFO - parrotllm.training - Val: loss=6.6243, ppl=753.17
255
+ 2026-04-08 10:25:04 - INFO - parrotllm.training - ** New best validation loss! **
256
+ 2026-04-08 10:25:05 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p6243_epoch_0002_step_0000170.pt
257
+ 2026-04-08 10:25:05 - INFO - parrotllm.training - ------------------------------------------------------------
258
+ 2026-04-08 10:25:07 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0002_step_0000170.pt
259
+ 2026-04-08 10:25:08 - INFO - parrotllm.training - Starting evaluation...
260
+ 2026-04-08 10:25:08 - INFO - parrotllm.training - ------------------------------------------------------------
261
+ 2026-04-08 10:25:08 - INFO - parrotllm.training - Epoch 3 complete:
262
+ 2026-04-08 10:25:08 - INFO - parrotllm.training - Train: loss=6.0717, ppl=433.40
263
+ 2026-04-08 10:25:08 - INFO - parrotllm.training - Val: loss=6.6202, ppl=750.06
264
+ 2026-04-08 10:25:08 - INFO - parrotllm.training - ** New best validation loss! **
265
+ 2026-04-08 10:25:08 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p6202_epoch_0003_step_0000174.pt
266
+ 2026-04-08 10:25:09 - INFO - parrotllm.training - ------------------------------------------------------------
267
+ 2026-04-08 10:25:10 - INFO - parrotllm.training - step 180 | epoch 3 | loss 5.9077 | lr 1.00e-03 | grad 0.7035
268
+ 2026-04-08 10:25:10 - INFO - parrotllm.training - Starting evaluation...
269
+ 2026-04-08 10:25:10 - INFO - parrotllm.training - ------------------------------------------------------------
270
+ 2026-04-08 10:25:10 - INFO - parrotllm.training - Train: loss=5.9077, ppl=367.86
271
+ 2026-04-08 10:25:10 - INFO - parrotllm.training - Val: loss=6.6468, ppl=770.28
272
+ 2026-04-08 10:25:10 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p6468_epoch_0003_step_0000180.pt
273
+ 2026-04-08 10:25:11 - INFO - parrotllm.training - No validation improvement for 1/3 evaluation(s) (best=6.6202, min_delta=0.000000).
274
+ 2026-04-08 10:25:11 - INFO - parrotllm.training - ------------------------------------------------------------
275
+ 2026-04-08 10:25:12 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0003_step_0000180.pt
276
+ 2026-04-08 10:25:14 - INFO - parrotllm.training - step 190 | epoch 3 | loss 5.9908 | lr 1.00e-03 | grad 0.6098
277
+ 2026-04-08 10:25:14 - INFO - parrotllm.training - Starting evaluation...
278
+ 2026-04-08 10:25:14 - INFO - parrotllm.training - ------------------------------------------------------------
279
+ 2026-04-08 10:25:14 - INFO - parrotllm.training - Train: loss=5.9908, ppl=399.72
280
+ 2026-04-08 10:25:14 - INFO - parrotllm.training - Val: loss=6.6222, ppl=751.60
281
+ 2026-04-08 10:25:16 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p6222_epoch_0003_step_0000190.pt
282
+ 2026-04-08 10:25:16 - INFO - parrotllm.training - No validation improvement for 2/3 evaluation(s) (best=6.6202, min_delta=0.000000).
283
+ 2026-04-08 10:25:16 - INFO - parrotllm.training - ------------------------------------------------------------
284
+ 2026-04-08 10:25:17 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0003_step_0000190.pt
285
+ 2026-04-08 10:25:19 - INFO - parrotllm.training - step 200 | epoch 3 | loss 5.9190 | lr 1.00e-03 | grad 0.5578
286
+ 2026-04-08 10:25:19 - INFO - parrotllm.training - Starting evaluation...
287
+ 2026-04-08 10:25:19 - INFO - parrotllm.training - ------------------------------------------------------------
288
+ 2026-04-08 10:25:19 - INFO - parrotllm.training - Train: loss=5.9190, ppl=372.04
289
+ 2026-04-08 10:25:19 - INFO - parrotllm.training - Val: loss=6.5935, ppl=730.32
290
+ 2026-04-08 10:25:19 - INFO - parrotllm.training - ** New best validation loss! **
291
+ 2026-04-08 10:25:19 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5935_epoch_0003_step_0000200.pt
292
+ 2026-04-08 10:25:20 - INFO - parrotllm.training - ------------------------------------------------------------
293
+ 2026-04-08 10:25:21 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0003_step_0000200.pt
294
+ 2026-04-08 10:25:23 - INFO - parrotllm.training - step 210 | epoch 3 | loss 5.8753 | lr 1.00e-03 | grad 0.6255
295
+ 2026-04-08 10:25:23 - INFO - parrotllm.training - Starting evaluation...
296
+ 2026-04-08 10:25:23 - INFO - parrotllm.training - ------------------------------------------------------------
297
+ 2026-04-08 10:25:23 - INFO - parrotllm.training - Train: loss=5.8753, ppl=356.14
298
+ 2026-04-08 10:25:23 - INFO - parrotllm.training - Val: loss=6.5643, ppl=709.30
299
+ 2026-04-08 10:25:23 - INFO - parrotllm.training - ** New best validation loss! **
300
+ 2026-04-08 10:25:24 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5643_epoch_0003_step_0000210.pt
301
+ 2026-04-08 10:25:24 - INFO - parrotllm.training - ------------------------------------------------------------
302
+ 2026-04-08 10:25:25 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0003_step_0000210.pt
303
+ 2026-04-08 10:25:27 - INFO - parrotllm.training - step 220 | epoch 3 | loss 5.9597 | lr 1.00e-03 | grad 0.6323
304
+ 2026-04-08 10:25:27 - INFO - parrotllm.training - Starting evaluation...
305
+ 2026-04-08 10:25:27 - INFO - parrotllm.training - ------------------------------------------------------------
306
+ 2026-04-08 10:25:27 - INFO - parrotllm.training - Train: loss=5.9597, ppl=387.51
307
+ 2026-04-08 10:25:27 - INFO - parrotllm.training - Val: loss=6.5386, ppl=691.29
308
+ 2026-04-08 10:25:27 - INFO - parrotllm.training - ** New best validation loss! **
309
+ 2026-04-08 10:25:27 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5386_epoch_0003_step_0000220.pt
310
+ 2026-04-08 10:25:28 - INFO - parrotllm.training - ------------------------------------------------------------
311
+ 2026-04-08 10:25:29 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0003_step_0000220.pt
312
+ 2026-04-08 10:25:31 - INFO - parrotllm.training - step 230 | epoch 3 | loss 5.9037 | lr 1.00e-03 | grad 0.5964
313
+ 2026-04-08 10:25:31 - INFO - parrotllm.training - Starting evaluation...
314
+ 2026-04-08 10:25:31 - INFO - parrotllm.training - ------------------------------------------------------------
315
+ 2026-04-08 10:25:31 - INFO - parrotllm.training - Train: loss=5.9037, ppl=366.38
316
+ 2026-04-08 10:25:31 - INFO - parrotllm.training - Val: loss=6.5273, ppl=683.53
317
+ 2026-04-08 10:25:31 - INFO - parrotllm.training - ** New best validation loss! **
318
+ 2026-04-08 10:25:33 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5273_epoch_0003_step_0000230.pt
319
+ 2026-04-08 10:25:33 - INFO - parrotllm.training - ------------------------------------------------------------
320
+ 2026-04-08 10:25:34 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0003_step_0000230.pt
321
+ 2026-04-08 10:25:34 - INFO - parrotllm.training - Starting evaluation...
322
+ 2026-04-08 10:25:34 - INFO - parrotllm.training - ------------------------------------------------------------
323
+ 2026-04-08 10:25:35 - INFO - parrotllm.training - Epoch 4 complete:
324
+ 2026-04-08 10:25:35 - INFO - parrotllm.training - Train: loss=5.9703, ppl=391.63
325
+ 2026-04-08 10:25:35 - INFO - parrotllm.training - Val: loss=6.5308, ppl=685.94
326
+ 2026-04-08 10:25:36 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5308_epoch_0004_step_0000232.pt
327
+ 2026-04-08 10:25:36 - INFO - parrotllm.training - No validation improvement for 1/3 evaluation(s) (best=6.5273, min_delta=0.000000).
328
+ 2026-04-08 10:25:36 - INFO - parrotllm.training - ------------------------------------------------------------
329
+ 2026-04-08 10:25:38 - INFO - parrotllm.training - step 240 | epoch 4 | loss 5.6591 | lr 1.00e-03 | grad 0.6297
330
+ 2026-04-08 10:25:38 - INFO - parrotllm.training - Starting evaluation...
331
+ 2026-04-08 10:25:38 - INFO - parrotllm.training - ------------------------------------------------------------
332
+ 2026-04-08 10:25:38 - INFO - parrotllm.training - Train: loss=5.6591, ppl=286.89
333
+ 2026-04-08 10:25:38 - INFO - parrotllm.training - Val: loss=6.5239, ppl=681.23
334
+ 2026-04-08 10:25:38 - INFO - parrotllm.training - ** New best validation loss! **
335
+ 2026-04-08 10:25:38 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5239_epoch_0004_step_0000240.pt
336
+ 2026-04-08 10:25:39 - INFO - parrotllm.training - ------------------------------------------------------------
337
+ 2026-04-08 10:25:40 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0004_step_0000240.pt
338
+ 2026-04-08 10:25:42 - INFO - parrotllm.training - step 250 | epoch 4 | loss 5.6930 | lr 8.50e-04 | grad 0.6920
339
+ 2026-04-08 10:25:42 - INFO - parrotllm.training - Starting evaluation...
340
+ 2026-04-08 10:25:42 - INFO - parrotllm.training - ------------------------------------------------------------
341
+ 2026-04-08 10:25:42 - INFO - parrotllm.training - Train: loss=5.6930, ppl=296.79
342
+ 2026-04-08 10:25:42 - INFO - parrotllm.training - Val: loss=6.5343, ppl=688.35
343
+ 2026-04-08 10:25:44 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5343_epoch_0004_step_0000250.pt
344
+ 2026-04-08 10:25:44 - INFO - parrotllm.training - No validation improvement for 1/3 evaluation(s) (best=6.5239, min_delta=0.000000).
345
+ 2026-04-08 10:25:44 - INFO - parrotllm.training - ------------------------------------------------------------
346
+ 2026-04-08 10:25:45 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0004_step_0000250.pt
347
+ 2026-04-08 10:25:47 - INFO - parrotllm.training - step 260 | epoch 4 | loss 5.5703 | lr 7.00e-04 | grad 0.5898
348
+ 2026-04-08 10:25:47 - INFO - parrotllm.training - Starting evaluation...
349
+ 2026-04-08 10:25:47 - INFO - parrotllm.training - ------------------------------------------------------------
350
+ 2026-04-08 10:25:47 - INFO - parrotllm.training - Train: loss=5.5703, ppl=262.50
351
+ 2026-04-08 10:25:47 - INFO - parrotllm.training - Val: loss=6.5350, ppl=688.82
352
+ 2026-04-08 10:25:48 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5350_epoch_0004_step_0000260.pt
353
+ 2026-04-08 10:25:48 - INFO - parrotllm.training - No validation improvement for 2/3 evaluation(s) (best=6.5239, min_delta=0.000000).
354
+ 2026-04-08 10:25:48 - INFO - parrotllm.training - ------------------------------------------------------------
355
+ 2026-04-08 10:25:50 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0004_step_0000260.pt
356
+ 2026-04-08 10:25:52 - INFO - parrotllm.training - step 270 | epoch 4 | loss 5.5189 | lr 5.50e-04 | grad 0.7256
357
+ 2026-04-08 10:25:52 - INFO - parrotllm.training - Starting evaluation...
358
+ 2026-04-08 10:25:52 - INFO - parrotllm.training - ------------------------------------------------------------
359
+ 2026-04-08 10:25:52 - INFO - parrotllm.training - Train: loss=5.5189, ppl=249.37
360
+ 2026-04-08 10:25:52 - INFO - parrotllm.training - Val: loss=6.5059, ppl=669.08
361
+ 2026-04-08 10:25:52 - INFO - parrotllm.training - ** New best validation loss! **
362
+ 2026-04-08 10:25:53 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5059_epoch_0004_step_0000270.pt
363
+ 2026-04-08 10:25:53 - INFO - parrotllm.training - ------------------------------------------------------------
364
+ 2026-04-08 10:25:54 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0004_step_0000270.pt
365
+ 2026-04-08 10:25:56 - INFO - parrotllm.training - step 280 | epoch 4 | loss 5.5293 | lr 4.00e-04 | grad 0.5655
366
+ 2026-04-08 10:25:56 - INFO - parrotllm.training - Starting evaluation...
367
+ 2026-04-08 10:25:56 - INFO - parrotllm.training - ------------------------------------------------------------
368
+ 2026-04-08 10:25:56 - INFO - parrotllm.training - Train: loss=5.5293, ppl=251.97
369
+ 2026-04-08 10:25:56 - INFO - parrotllm.training - Val: loss=6.5027, ppl=666.91
370
+ 2026-04-08 10:25:56 - INFO - parrotllm.training - ** New best validation loss! **
371
+ 2026-04-08 10:25:56 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5027_epoch_0004_step_0000280.pt
372
+ 2026-04-08 10:25:57 - INFO - parrotllm.training - ------------------------------------------------------------
373
+ 2026-04-08 10:25:57 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0004_step_0000280.pt
374
+ 2026-04-08 10:26:00 - INFO - parrotllm.training - step 290 | epoch 4 | loss 5.6288 | lr 2.50e-04 | grad 0.6188
375
+ 2026-04-08 10:26:00 - INFO - parrotllm.training - Starting evaluation...
376
+ 2026-04-08 10:26:00 - INFO - parrotllm.training - ------------------------------------------------------------
377
+ 2026-04-08 10:26:00 - INFO - parrotllm.training - Epoch 5 complete:
378
+ 2026-04-08 10:26:00 - INFO - parrotllm.training - Train: loss=5.6288, ppl=278.32
379
+ 2026-04-08 10:26:00 - INFO - parrotllm.training - Val: loss=6.4871, ppl=656.64
380
+ 2026-04-08 10:26:00 - INFO - parrotllm.training - ** New best validation loss! **
381
+ 2026-04-08 10:26:01 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p4871_epoch_0005_step_0000290.pt
382
+ 2026-04-08 10:26:01 - INFO - parrotllm.training - ------------------------------------------------------------
383
+ 2026-04-08 10:26:02 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0005_step_0000290.pt
384
+ 2026-04-08 10:26:05 - INFO - parrotllm.training - step 300 | epoch 5 | loss 5.3922 | lr 1.00e-04 | grad 0.4955
385
+ 2026-04-08 10:26:05 - INFO - parrotllm.training - Starting evaluation...
386
+ 2026-04-08 10:26:05 - INFO - parrotllm.training - ------------------------------------------------------------
387
+ 2026-04-08 10:26:05 - INFO - parrotllm.training - Train: loss=5.3922, ppl=219.68
388
+ 2026-04-08 10:26:05 - INFO - parrotllm.training - Val: loss=6.5104, ppl=672.12
389
+ 2026-04-08 10:26:05 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/best_loss_6p5104_epoch_0005_step_0000300.pt
390
+ 2026-04-08 10:26:06 - INFO - parrotllm.training - No validation improvement for 1/3 evaluation(s) (best=6.4871, min_delta=0.000000).
391
+ 2026-04-08 10:26:06 - INFO - parrotllm.training - ------------------------------------------------------------
392
+ 2026-04-08 10:26:07 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0005_step_0000300.pt
393
+ 2026-04-08 10:26:09 - INFO - parrotllm.training - Saved final checkpoint: /home/user/code/ParrotLLM/runs/fast_learn/run_20260408_102350/checkpoints/last_epoch_0005_step_0000300.pt
394
+ 2026-04-08 10:26:09 - INFO - parrotllm.training -
395
+ Loss Curve:
396
+ Loss
397
+ ^
398
+ 10.88 |*
399
+ |**
400
+ | **
401
+ | *
402
+ | **
403
+ | *******
404
+ | *******************
405
+ 5.32 | **********************
406
+ +--------------------------------------------------> step
407
+ 1 300
408
+ 2026-04-08 10:26:09 - INFO - parrotllm.training -
409
+ ============================================================
410
+ TRAINING COMPLETE
411
+ ============================================================
412
+ Epochs: 6
413
+ Total steps: 300
414
+ Total time: 0.04 hours
415
+ Best validation loss: 6.4871
416
+ Run directory: runs/fast_learn/run_20260408_102350
417
+ ============================================================
418
+ 2026-04-08 10:26:09 - INFO - parrotllm.training - Uploading completed run directory to Hugging Face: repo=ParrotLabs/Preprocessed
runs/fast_learn/run_20260408_102350/training_plots.pdf ADDED
Binary file (33.3 kB). View file