Hugging Face's logo

Datasets:
ParrotLabs
/
Preprocessed

Dataset card Files
xet
 Community
thaferbe commited on
Commit
cf4af51
·
verified ·
1 Parent(s): 04d6525

Upload training run run_20260408_135126

Browse files

ParrotLLM end-of-training sync for local run runs/smoketest/exp_c/run_20260408_135126.

Files changed (24) hide show
  1. runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p7414_epoch_0004_step_0000012.pt +3 -0
  2. runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p7523_epoch_0003_step_0000011.pt +3 -0
  3. runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p8117_epoch_0003_step_0000010.pt +3 -0
  4. runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p9587_epoch_0003_step_0000009.pt +3 -0
  5. runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p1591_epoch_0002_step_0000008.pt +3 -0
  6. runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p4107_epoch_0002_step_0000007.pt +3 -0
  7. runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p7135_epoch_0002_step_0000006.pt +3 -0
  8. runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p0758_epoch_0001_step_0000005.pt +3 -0
  9. runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p4788_epoch_0001_step_0000004.pt +3 -0
  10. runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p9060_epoch_0001_step_0000003.pt +3 -0
  11. runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000003.pt +3 -0
  12. runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000004.pt +3 -0
  13. runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000005.pt +3 -0
  14. runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000006.pt +3 -0
  15. runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000007.pt +3 -0
  16. runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000008.pt +3 -0
  17. runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000009.pt +3 -0
  18. runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000010.pt +3 -0
  19. runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000011.pt +3 -0
  20. runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0004_step_0000012.pt +3 -0
  21. runs/smoketest/exp_c/run_20260408_135126/config.json +70 -0
  22. runs/smoketest/exp_c/run_20260408_135126/metrics.jsonl +54 -0
  23. runs/smoketest/exp_c/run_20260408_135126/train.log +297 -0
  24. runs/smoketest/exp_c/run_20260408_135126/training_plots.pdf +0 -0
runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p7414_epoch_0004_step_0000012.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d1fd07f831d30d822183635ebccfd2b0274b013794dcaac95313a5e9cb32e761
3
+ size 479869180
runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p7523_epoch_0003_step_0000011.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b3cedd9a05dd38be077fef8a7121241ef34a29f63e9ba4f2eaee988150b4c05d
3
+ size 479869180
runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p8117_epoch_0003_step_0000010.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:4bc89010af7ad27c73a9e03decab1aa05bdcb7ad227808e696854f37fd10825c
3
+ size 479869180
runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p9587_epoch_0003_step_0000009.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:61f9905e6b722773dc834c0dafdc82813e3a32ea57a270613d8cb08e1c7fbe46
3
+ size 479869180
runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p1591_epoch_0002_step_0000008.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:fc458a3aed67a58b450d85b1234a3ec53f3e3cc38bf250cfc6f08df87b32e95c
3
+ size 479869180
runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p4107_epoch_0002_step_0000007.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5545a1875a8c6581cd846374c9e148ca1ff1596a65e6bcabafbc8efe6b6b70cb
3
+ size 479869180
runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p7135_epoch_0002_step_0000006.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:396e7f890693ff4c9e66c5a032b7f64f7b9726e2444f8905fba4dcd271a73ae1
3
+ size 479869180
runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p0758_epoch_0001_step_0000005.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:521c483b0b3d397bb36de36dc67ea70e889a21e03acb7556511d72a73c630a8e
3
+ size 479869180
runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p4788_epoch_0001_step_0000004.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:901f7c489858f42284eb3901fedc5d709dfe2b5294beeaf360d0c5c5df1dfde2
3
+ size 479869180
runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p9060_epoch_0001_step_0000003.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:04cd233af6ddea22f6761b3b496d97e9d3152f4e52bd9c3df29531442f441bac
3
+ size 479869180
runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000003.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5f491db31011ef9897109678c69da97f8eee3294c21b2a4c52ac577402e41842
3
+ size 479860168
runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000004.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:931e2b4f715ce6d6bdec3c0b5bc392aefac7157bf9f97e1ae9b0dbb32a609ab2
3
+ size 479860168
runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000005.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:10459c8feacc86abc3820d50a906ec37266fb2f997651fd9f80e5332b95df9dd
3
+ size 479860168
runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000006.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:f25e2f0e4bc78191bcc51a49aa69328f49d9d583cf7912fc5590199fb9ba123b
3
+ size 479860168
runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000007.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e5f9f52b072b46c6dfd87ef63a7be12ae88b5b5ded5bcfae4c1fd03c05847ea
3
+ size 479860168
runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000008.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:45b8a2f9756b7c1dad12a591a6edf917adacb9adc51664629df6c3ffce4ee03f
3
+ size 479860168
runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000009.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:49c66aaf2a8758cc0401d66b4f50495422f2dd9683c601dc7111df975467a3b2
3
+ size 479860168
runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000010.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c305b0a5291901cf5b5abd7e37325a44af0fb60819d8b4d0028dfa59e070c49a
3
+ size 479860168
runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000011.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:54c096fc1c02c4b282e20b65eb80fb69fad150a06f5a1e1293e69f47d22ec88e
3
+ size 479860168
runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0004_step_0000012.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:9a38f2944e656acbbbbf4e9929ea20b5a943067f89b617badee03a5e3327bc99
3
+ size 479860168
runs/smoketest/exp_c/run_20260408_135126/config.json ADDED
@@ -0,0 +1,70 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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": 384,
9
+ "n_layers": 14,
10
+ "n_heads": 6,
11
+ "d_ff": 768,
12
+ "context_length": 1024,
13
+ "bias": false,
14
+ "dropout": 0.005,
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": 4,
23
+ "pin_memory": true,
24
+ "batch_size": 32,
25
+ "gradient_accumulation_steps": 1,
26
+ "train_samples_per_epoch": null,
27
+ "val_sequence_stride": null,
28
+ "learning_rate": 0.001305,
29
+ "min_lr": 0.0001305,
30
+ "weight_decay": 0.063,
31
+ "beta1": 0.9,
32
+ "beta2": 0.9978,
33
+ "grad_clip": 1.786,
34
+ "warmup_steps": 4,
35
+ "max_steps": 12,
36
+ "lr_schedule": "wsd",
37
+ "lr_decay_ratio": 0.179,
38
+ "z_loss_coeff": 0.000275,
39
+ "save_every": 1,
40
+ "eval_every": 1,
41
+ "checkpoint_dir": "checkpoints",
42
+ "keep_last_checkpoints": 10,
43
+ "keep_best_checkpoints": 10,
44
+ "early_stopping_patience": 0,
45
+ "early_stopping_min_delta": 0.001,
46
+ "runs_dir": "runs/smoketest/exp_c",
47
+ "log_every": 1,
48
+ "hf_upload": {
49
+ "repo_id": "ParrotLabs/Preprocessed",
50
+ "repo_type": "dataset",
51
+ "path_in_repo": "",
52
+ "private": null
53
+ },
54
+ "compile": true
55
+ },
56
+ "tune": null,
57
+ "logging": {
58
+ "console_level": "INFO",
59
+ "file_level": "DEBUG",
60
+ "components": {
61
+ "data_preprocessing": "DEBUG",
62
+ "model_initialization": "INFO",
63
+ "training": "INFO"
64
+ },
65
+ "profiler": null
66
+ },
67
+ "eval": null,
68
+ "inference": null,
69
+ "chat": null
70
+ }
runs/smoketest/exp_c/run_20260408_135126/metrics.jsonl ADDED
@@ -0,0 +1,54 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {"stage": "pretraining", "type": "config", "timestamp": "2026-04-08T13:51:27.157589+00:00", "device": "auto", "train_bin": "data/SmokeTest/train.bin", "val_bin": "data/SmokeTest/val.bin", "num_workers": 4, "pin_memory": true, "batch_size": 32, "gradient_accumulation_steps": 1, "train_samples_per_epoch": null, "val_sequence_stride": null, "learning_rate": 0.001305, "min_lr": 0.0001305, "weight_decay": 0.063, "beta1": 0.9, "beta2": 0.9978, "grad_clip": 1.786, "warmup_steps": 4, "max_steps": 12, "lr_schedule": "wsd", "lr_decay_ratio": 0.179, "z_loss_coeff": 0.000275, "save_every": 1, "eval_every": 1, "checkpoint_dir": "checkpoints", "keep_last_checkpoints": 10, "keep_best_checkpoints": 10, "early_stopping_patience": 0, "early_stopping_min_delta": 0.001, "runs_dir": "runs/smoketest/exp_c", "log_every": 1, "hf_upload": {"repo_id": "ParrotLabs/Preprocessed", "repo_type": "dataset", "path_in_repo": "", "private": null}, "compile": true}
2
+ {"stage": "pretraining", "type": "checkpoint_policy", "timestamp": "2026-04-08T13:51:27.157675+00:00", "checkpoint_dir": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints", "keep_last": 10, "keep_best": 10}
3
+ {"stage": "pretraining", "type": "model_architecture", "timestamp": "2026-04-08T13:51:28.981286+00:00", "vocab_size": 50258, "context_length": 1024, "n_layers": 14, "n_heads": 6, "d_model": 384, "d_ff": 768, "dropout": 0.005, "bias": false, "total_params": 39966592, "total_params_non_embedding": 39966592, "trainable_params": 39966592, "non_trainable_params": 0, "params_size_mb": 159.87}
4
+ {"stage": "pretraining", "type": "initial_validation", "timestamp": "2026-04-08T13:51:34.747315+00:00", "val_loss": 10.906839370727539, "val_ppl": 54548.16603071756}
5
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:51:58.612750+00:00", "epoch": 0, "step": 1, "train_loss": 10.931233406066895, "perplexity": 55895.17866096571, "lr": 0.00032625, "grad_norm": 6.673876, "tokens_per_sec": 1373}
6
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:52:01.225851+00:00", "step": 1, "epoch": 0, "val_loss": 10.584296703338623, "val_ppl": 39509.51086658796, "eval_train_loss": 10.931233406066895, "eval_train_ppl": 55895.17866096571}
7
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:52:02.210789+00:00", "step": 1, "epoch": 0, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_10p5843_epoch_0000_step_0000001.pt", "val_loss": 10.584296703338623}
8
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:52:04.413321+00:00", "step": 1, "epoch": 0, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0000_step_0000001.pt", "category": "last"}
9
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:52:05.316901+00:00", "epoch": 0, "step": 2, "train_loss": 10.6100492477417, "perplexity": 40540.19571687804, "lr": 0.0006525, "grad_norm": 15.277437, "tokens_per_sec": 4888}
10
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:52:07.913631+00:00", "step": 2, "epoch": 0, "val_loss": 10.239965915679932, "val_ppl": 28000.17154315802, "eval_train_loss": 10.6100492477417, "eval_train_ppl": 40540.19571687804}
11
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:52:10.459586+00:00", "step": 2, "epoch": 0, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_10p2400_epoch_0000_step_0000002.pt", "val_loss": 10.239965915679932}
12
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:52:13.009873+00:00", "step": 2, "epoch": 0, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0000_step_0000002.pt", "category": "last"}
13
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:52:13.951572+00:00", "epoch": 0, "step": 3, "train_loss": 10.26486587524414, "perplexity": 28706.12734333692, "lr": 0.00097875, "grad_norm": 1.781714, "tokens_per_sec": 3805}
14
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:52:16.553201+00:00", "step": 3, "epoch": 1, "val_loss": 9.905978202819824, "val_ppl": 20049.875090595524, "eval_train_loss": 10.26486587524414, "eval_train_ppl": 28706.12734333692}
15
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:52:17.855075+00:00", "step": 3, "epoch": 1, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p9060_epoch_0001_step_0000003.pt", "val_loss": 9.905978202819824}
16
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:52:19.598854+00:00", "step": 3, "epoch": 1, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000003.pt", "category": "last"}
17
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:52:20.787965+00:00", "epoch": 1, "step": 4, "train_loss": 9.911581039428711, "perplexity": 20162.526553934145, "lr": 0.001305, "grad_norm": 1.649637, "tokens_per_sec": 4778}
18
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:52:23.352056+00:00", "step": 4, "epoch": 1, "val_loss": 9.478840827941895, "val_ppl": 13080.015734343084, "eval_train_loss": 9.911581039428711, "eval_train_ppl": 20162.526553934145}
19
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:52:25.969581+00:00", "step": 4, "epoch": 1, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p4788_epoch_0001_step_0000004.pt", "val_loss": 9.478840827941895}
20
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:52:28.855939+00:00", "step": 4, "epoch": 1, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000004.pt", "category": "last"}
21
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:52:29.787595+00:00", "epoch": 1, "step": 5, "train_loss": 9.487369537353516, "perplexity": 13192.048455809945, "lr": 0.001305, "grad_norm": 1.59308, "tokens_per_sec": 3641}
22
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:52:32.387285+00:00", "step": 5, "epoch": 1, "val_loss": 9.075779438018799, "val_ppl": 8740.996149051123, "eval_train_loss": 9.487369537353516, "eval_train_ppl": 13192.048455809945}
23
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:52:33.922187+00:00", "step": 5, "epoch": 1, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p0758_epoch_0001_step_0000005.pt", "val_loss": 9.075779438018799}
24
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:52:36.834695+00:00", "step": 5, "epoch": 1, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000005.pt", "category": "last"}
25
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:52:37.810246+00:00", "epoch": 1, "step": 6, "train_loss": 9.075180053710938, "perplexity": 8735.758502962552, "lr": 0.001305, "grad_norm": 1.57253, "tokens_per_sec": 4084}
26
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:52:40.403789+00:00", "step": 6, "epoch": 2, "val_loss": 8.713451385498047, "val_ppl": 6084.2052296725615, "eval_train_loss": 9.075180053710938, "eval_train_ppl": 8735.758502962552}
27
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:52:43.002912+00:00", "step": 6, "epoch": 2, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p7135_epoch_0002_step_0000006.pt", "val_loss": 8.713451385498047}
28
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:52:46.265905+00:00", "step": 6, "epoch": 2, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000006.pt", "category": "last"}
29
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:52:47.211385+00:00", "epoch": 2, "step": 7, "train_loss": 8.709702491760254, "perplexity": 6061.438891742002, "lr": 0.001305, "grad_norm": 1.517315, "tokens_per_sec": 3486}
30
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:52:49.823718+00:00", "step": 7, "epoch": 2, "val_loss": 8.410696983337402, "val_ppl": 4494.892285051032, "eval_train_loss": 8.709702491760254, "eval_train_ppl": 6061.438891742002}
31
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:52:51.149996+00:00", "step": 7, "epoch": 2, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p4107_epoch_0002_step_0000007.pt", "val_loss": 8.410696983337402}
32
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:52:52.782409+00:00", "step": 7, "epoch": 2, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000007.pt", "category": "last"}
33
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:52:53.672340+00:00", "epoch": 2, "step": 8, "train_loss": 8.431455612182617, "perplexity": 4589.17529299602, "lr": 0.001305, "grad_norm": 1.458246, "tokens_per_sec": 5072}
34
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:52:56.265052+00:00", "step": 8, "epoch": 2, "val_loss": 8.159146785736084, "val_ppl": 3495.2031739885692, "eval_train_loss": 8.431455612182617, "eval_train_ppl": 4589.17529299602}
35
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:52:58.622110+00:00", "step": 8, "epoch": 2, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p1591_epoch_0002_step_0000008.pt", "val_loss": 8.159146785736084}
36
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:53:01.622114+00:00", "step": 8, "epoch": 2, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000008.pt", "category": "last"}
37
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:53:02.521759+00:00", "epoch": 2, "step": 9, "train_loss": 8.140697479248047, "perplexity": 3431.31030128376, "lr": 0.001305, "grad_norm": 1.334799, "tokens_per_sec": 3703}
38
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:53:05.114807+00:00", "step": 9, "epoch": 3, "val_loss": 7.958655834197998, "val_ppl": 2860.225749807668, "eval_train_loss": 8.140697479248047, "eval_train_ppl": 3431.31030128376}
39
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:53:06.712384+00:00", "step": 9, "epoch": 3, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p9587_epoch_0003_step_0000009.pt", "val_loss": 7.958655834197998}
40
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:53:10.160461+00:00", "step": 9, "epoch": 3, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000009.pt", "category": "last"}
41
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:53:11.429770+00:00", "epoch": 3, "step": 10, "train_loss": 7.9377264976501465, "perplexity": 2800.9852174535044, "lr": 0.001305, "grad_norm": 1.148983, "tokens_per_sec": 3678}
42
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:53:14.678601+00:00", "step": 10, "epoch": 3, "val_loss": 7.811683893203735, "val_ppl": 2469.28494856854, "eval_train_loss": 7.9377264976501465, "eval_train_ppl": 2800.9852174535044}
43
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:53:16.458566+00:00", "step": 10, "epoch": 3, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p8117_epoch_0003_step_0000010.pt", "val_loss": 7.811683893203735}
44
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:53:19.457258+00:00", "step": 10, "epoch": 3, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000010.pt", "category": "last"}
45
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:53:20.339018+00:00", "epoch": 3, "step": 11, "train_loss": 7.7149338722229, "perplexity": 2241.5746625317774, "lr": 0.00071775, "grad_norm": 0.894213, "tokens_per_sec": 3678}
46
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:53:22.954115+00:00", "step": 11, "epoch": 3, "val_loss": 7.7522807121276855, "val_ppl": 2326.8732955663254, "eval_train_loss": 7.7149338722229, "eval_train_ppl": 2241.5746625317774}
47
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:53:24.337636+00:00", "step": 11, "epoch": 3, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p7523_epoch_0003_step_0000011.pt", "val_loss": 7.7522807121276855}
48
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:53:25.999610+00:00", "step": 11, "epoch": 3, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000011.pt", "category": "last"}
49
+ {"stage": "pretraining", "type": "step", "timestamp": "2026-04-08T13:53:26.877858+00:00", "epoch": 3, "step": 12, "train_loss": 7.639630317687988, "perplexity": 2078.975114165463, "lr": 0.00013049999999999997, "grad_norm": 0.755258, "tokens_per_sec": 5011}
50
+ {"stage": "pretraining", "type": "eval", "timestamp": "2026-04-08T13:53:29.473263+00:00", "step": 12, "epoch": 4, "val_loss": 7.741426229476929, "val_ppl": 2301.7528709952203, "eval_train_loss": 7.639630317687988, "eval_train_ppl": 2078.975114165463}
51
+ {"stage": "pretraining", "type": "best_checkpoint", "timestamp": "2026-04-08T13:53:32.002917+00:00", "step": 12, "epoch": 4, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p7414_epoch_0004_step_0000012.pt", "val_loss": 7.741426229476929}
52
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:53:35.430576+00:00", "step": 12, "epoch": 4, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0004_step_0000012.pt", "category": "last"}
53
+ {"stage": "pretraining", "type": "checkpoint", "timestamp": "2026-04-08T13:53:37.772983+00:00", "step": 12, "epoch": 4, "path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0004_step_0000012.pt", "category": "last"}
54
+ {"stage": "pretraining", "type": "training_complete", "timestamp": "2026-04-08T13:53:37.773276+00:00", "epochs": 4, "total_steps": 12, "total_time_hours": 0.03, "best_val_loss": 7.741426229476929, "best_checkpoint_path": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p7414_epoch_0004_step_0000012.pt", "checkpoint_dir": "/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints", "stopped_early": false, "stop_reason": null, "run_dir": "runs/smoketest/exp_c/run_20260408_135126"}
runs/smoketest/exp_c/run_20260408_135126/train.log ADDED
@@ -0,0 +1,297 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ 2026-04-08 13:51:27 - INFO - parrotllm - Logging initialised -> runs/smoketest/exp_c/run_20260408_135126/train.log
2
+ 2026-04-08 13:51:27 - INFO - parrotllm.training - device=cuda:0 | rank=0 | world_size=8 | distributed=yes
3
+ 2026-04-08 13:51:27 - INFO - parrotllm.training - checkpoint directory=/home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints | keep_last=10 | keep_best=10
4
+ 2026-04-08 13:51:27 - INFO - parrotllm.training - train windows=random-overlapping | tokens=947045 | valid_starts=946021 | sampled_windows_per_epoch=923
5
+ 2026-04-08 13:51:27 - INFO - parrotllm.training - val windows=sequential | tokens=49601 | stride=1024 | windows=48
6
+ 2026-04-08 13:51:27 - INFO - parrotllm.training - compiling model with torch.compile...
7
+ 2026-04-08 13:51:28 - INFO - parrotllm.training -
8
+ ============================================================
9
+ MODEL ARCHITECTURE SUMMARY
10
+ ============================================================
11
+
12
+ Configuration:
13
+ Vocab size: 50258
14
+ Block size (context): 1024
15
+ Layers: 14
16
+ Heads: 6
17
+ Embedding dim: 384
18
+ FFN hidden dim: 768
19
+ Dropout: 0.005
20
+ Bias: False
21
+
22
+ Parameters (unique, weight-tied layers counted once):
23
+ Total: 39,966,592 (39.97M)
24
+ Trainable: 39,966,592 (39.97M)
25
+ Non-trainable: 0
26
+ Non-embedding: 39,966,592 (39.97M)
27
+ Position embeddings: 0 (RoPE: 0 learned params)
28
+ Size (MB): 159.87
29
+
30
+ Layer-wise breakdown (torchinfo):
31
+ Note: torchinfo double-counts weight-tied layers (tok_emb/lm_head).
32
+ ============================================================================================================================================
33
+ Layer (type:depth-idx) Input Shape Output Shape Param # Trainable
34
+ ============================================================================================================================================
35
+ ParrotLLM [32, 1024] [32, 1024, 50258] -- True
36
+ ├─Embedding: 1-1 [32, 1024] [32, 1024, 384] 19,299,072 True
37
+ ├─Dropout: 1-2 [32, 1024, 384] [32, 1024, 384] -- --
38
+ ├─ModuleList: 1-3 -- -- -- True
39
+ │ └─TransformerBlock: 2-1 [32, 1024, 384] [32, 1024, 384] -- True
40
+ │ │ └─RMSNorm: 3-1 [32, 1024, 384] [32, 1024, 384] 384 True
41
+ │ │ └─MultiHeadAttention: 3-2 [32, 1024, 384] [32, 1024, 384] 589,952 True
42
+ │ │ └─RMSNorm: 3-3 [32, 1024, 384] [32, 1024, 384] 384 True
43
+ │ │ └─RMSNorm: 3-4 [32, 1024, 384] [32, 1024, 384] 384 True
44
+ │ │ └─SwiGLUMLP: 3-5 [32, 1024, 384] [32, 1024, 384] 884,736 True
45
+ │ │ └─RMSNorm: 3-6 [32, 1024, 384] [32, 1024, 384] 384 True
46
+ │ └─TransformerBlock: 2-2 [32, 1024, 384] [32, 1024, 384] -- True
47
+ │ │ └─RMSNorm: 3-7 [32, 1024, 384] [32, 1024, 384] 384 True
48
+ │ │ └─MultiHeadAttention: 3-8 [32, 1024, 384] [32, 1024, 384] 589,952 True
49
+ │ │ └─RMSNorm: 3-9 [32, 1024, 384] [32, 1024, 384] 384 True
50
+ │ │ └─RMSNorm: 3-10 [32, 1024, 384] [32, 1024, 384] 384 True
51
+ │ │ └─SwiGLUMLP: 3-11 [32, 1024, 384] [32, 1024, 384] 884,736 True
52
+ │ │ └─RMSNorm: 3-12 [32, 1024, 384] [32, 1024, 384] 384 True
53
+ │ └─TransformerBlock: 2-3 [32, 1024, 384] [32, 1024, 384] -- True
54
+ │ │ └─RMSNorm: 3-13 [32, 1024, 384] [32, 1024, 384] 384 True
55
+ │ │ └─MultiHeadAttention: 3-14 [32, 1024, 384] [32, 1024, 384] 589,952 True
56
+ │ │ └─RMSNorm: 3-15 [32, 1024, 384] [32, 1024, 384] 384 True
57
+ │ │ └─RMSNorm: 3-16 [32, 1024, 384] [32, 1024, 384] 384 True
58
+ │ │ └─SwiGLUMLP: 3-17 [32, 1024, 384] [32, 1024, 384] 884,736 True
59
+ │ │ └─RMSNorm: 3-18 [32, 1024, 384] [32, 1024, 384] 384 True
60
+ │ └─TransformerBlock: 2-4 [32, 1024, 384] [32, 1024, 384] -- True
61
+ │ │ └─RMSNorm: 3-19 [32, 1024, 384] [32, 1024, 384] 384 True
62
+ │ │ └─MultiHeadAttention: 3-20 [32, 1024, 384] [32, 1024, 384] 589,952 True
63
+ │ │ └─RMSNorm: 3-21 [32, 1024, 384] [32, 1024, 384] 384 True
64
+ │ │ └─RMSNorm: 3-22 [32, 1024, 384] [32, 1024, 384] 384 True
65
+ │ │ └─SwiGLUMLP: 3-23 [32, 1024, 384] [32, 1024, 384] 884,736 True
66
+ │ │ └─RMSNorm: 3-24 [32, 1024, 384] [32, 1024, 384] 384 True
67
+ │ └─TransformerBlock: 2-5 [32, 1024, 384] [32, 1024, 384] -- True
68
+ │ │ └─RMSNorm: 3-25 [32, 1024, 384] [32, 1024, 384] 384 True
69
+ │ │ └─MultiHeadAttention: 3-26 [32, 1024, 384] [32, 1024, 384] 589,952 True
70
+ │ │ └─RMSNorm: 3-27 [32, 1024, 384] [32, 1024, 384] 384 True
71
+ │ │ └─RMSNorm: 3-28 [32, 1024, 384] [32, 1024, 384] 384 True
72
+ │ │ └─SwiGLUMLP: 3-29 [32, 1024, 384] [32, 1024, 384] 884,736 True
73
+ │ │ └─RMSNorm: 3-30 [32, 1024, 384] [32, 1024, 384] 384 True
74
+ │ └─TransformerBlock: 2-6 [32, 1024, 384] [32, 1024, 384] -- True
75
+ │ │ └─RMSNorm: 3-31 [32, 1024, 384] [32, 1024, 384] 384 True
76
+ │ │ └─MultiHeadAttention: 3-32 [32, 1024, 384] [32, 1024, 384] 589,952 True
77
+ │ │ └─RMSNorm: 3-33 [32, 1024, 384] [32, 1024, 384] 384 True
78
+ │ │ └─RMSNorm: 3-34 [32, 1024, 384] [32, 1024, 384] 384 True
79
+ │ │ └─SwiGLUMLP: 3-35 [32, 1024, 384] [32, 1024, 384] 884,736 True
80
+ │ │ └─RMSNorm: 3-36 [32, 1024, 384] [32, 1024, 384] 384 True
81
+ │ └─TransformerBlock: 2-7 [32, 1024, 384] [32, 1024, 384] -- True
82
+ │ │ └─RMSNorm: 3-37 [32, 1024, 384] [32, 1024, 384] 384 True
83
+ │ │ └─MultiHeadAttention: 3-38 [32, 1024, 384] [32, 1024, 384] 589,952 True
84
+ │ │ └─RMSNorm: 3-39 [32, 1024, 384] [32, 1024, 384] 384 True
85
+ │ │ └─RMSNorm: 3-40 [32, 1024, 384] [32, 1024, 384] 384 True
86
+ │ │ └─SwiGLUMLP: 3-41 [32, 1024, 384] [32, 1024, 384] 884,736 True
87
+ │ │ └─RMSNorm: 3-42 [32, 1024, 384] [32, 1024, 384] 384 True
88
+ │ └─TransformerBlock: 2-8 [32, 1024, 384] [32, 1024, 384] -- True
89
+ │ │ └─RMSNorm: 3-43 [32, 1024, 384] [32, 1024, 384] 384 True
90
+ │ │ └─MultiHeadAttention: 3-44 [32, 1024, 384] [32, 1024, 384] 589,952 True
91
+ │ │ └─RMSNorm: 3-45 [32, 1024, 384] [32, 1024, 384] 384 True
92
+ │ │ └─RMSNorm: 3-46 [32, 1024, 384] [32, 1024, 384] 384 True
93
+ │ │ └─SwiGLUMLP: 3-47 [32, 1024, 384] [32, 1024, 384] 884,736 True
94
+ │ │ └─RMSNorm: 3-48 [32, 1024, 384] [32, 1024, 384] 384 True
95
+ │ └─TransformerBlock: 2-9 [32, 1024, 384] [32, 1024, 384] -- True
96
+ │ │ └─RMSNorm: 3-49 [32, 1024, 384] [32, 1024, 384] 384 True
97
+ │ │ └─MultiHeadAttention: 3-50 [32, 1024, 384] [32, 1024, 384] 589,952 True
98
+ │ │ └─RMSNorm: 3-51 [32, 1024, 384] [32, 1024, 384] 384 True
99
+ │ │ └─RMSNorm: 3-52 [32, 1024, 384] [32, 1024, 384] 384 True
100
+ │ │ └─SwiGLUMLP: 3-53 [32, 1024, 384] [32, 1024, 384] 884,736 True
101
+ │ │ └─RMSNorm: 3-54 [32, 1024, 384] [32, 1024, 384] 384 True
102
+ │ └─TransformerBlock: 2-10 [32, 1024, 384] [32, 1024, 384] -- True
103
+ │ │ └─RMSNorm: 3-55 [32, 1024, 384] [32, 1024, 384] 384 True
104
+ │ │ └─MultiHeadAttention: 3-56 [32, 1024, 384] [32, 1024, 384] 589,952 True
105
+ │ │ └─RMSNorm: 3-57 [32, 1024, 384] [32, 1024, 384] 384 True
106
+ │ │ └─RMSNorm: 3-58 [32, 1024, 384] [32, 1024, 384] 384 True
107
+ │ │ └─SwiGLUMLP: 3-59 [32, 1024, 384] [32, 1024, 384] 884,736 True
108
+ │ │ └─RMSNorm: 3-60 [32, 1024, 384] [32, 1024, 384] 384 True
109
+ │ └─TransformerBlock: 2-11 [32, 1024, 384] [32, 1024, 384] -- True
110
+ │ │ └─RMSNorm: 3-61 [32, 1024, 384] [32, 1024, 384] 384 True
111
+ │ │ └─MultiHeadAttention: 3-62 [32, 1024, 384] [32, 1024, 384] 589,952 True
112
+ │ │ └─RMSNorm: 3-63 [32, 1024, 384] [32, 1024, 384] 384 True
113
+ │ │ └─RMSNorm: 3-64 [32, 1024, 384] [32, 1024, 384] 384 True
114
+ │ │ └─SwiGLUMLP: 3-65 [32, 1024, 384] [32, 1024, 384] 884,736 True
115
+ │ │ └─RMSNorm: 3-66 [32, 1024, 384] [32, 1024, 384] 384 True
116
+ │ └─TransformerBlock: 2-12 [32, 1024, 384] [32, 1024, 384] -- True
117
+ │ │ └─RMSNorm: 3-67 [32, 1024, 384] [32, 1024, 384] 384 True
118
+ │ │ └─MultiHeadAttention: 3-68 [32, 1024, 384] [32, 1024, 384] 589,952 True
119
+ │ │ └─RMSNorm: 3-69 [32, 1024, 384] [32, 1024, 384] 384 True
120
+ │ │ └─RMSNorm: 3-70 [32, 1024, 384] [32, 1024, 384] 384 True
121
+ │ │ └─SwiGLUMLP: 3-71 [32, 1024, 384] [32, 1024, 384] 884,736 True
122
+ │ │ └─RMSNorm: 3-72 [32, 1024, 384] [32, 1024, 384] 384 True
123
+ │ └─TransformerBlock: 2-13 [32, 1024, 384] [32, 1024, 384] -- True
124
+ │ │ └─RMSNorm: 3-73 [32, 1024, 384] [32, 1024, 384] 384 True
125
+ │ │ └─MultiHeadAttention: 3-74 [32, 1024, 384] [32, 1024, 384] 589,952 True
126
+ │ │ └─RMSNorm: 3-75 [32, 1024, 384] [32, 1024, 384] 384 True
127
+ │ │ └─RMSNorm: 3-76 [32, 1024, 384] [32, 1024, 384] 384 True
128
+ │ │ └─SwiGLUMLP: 3-77 [32, 1024, 384] [32, 1024, 384] 884,736 True
129
+ │ │ └─RMSNorm: 3-78 [32, 1024, 384] [32, 1024, 384] 384 True
130
+ │ └─TransformerBlock: 2-14 [32, 1024, 384] [32, 1024, 384] -- True
131
+ │ │ └─RMSNorm: 3-79 [32, 1024, 384] [32, 1024, 384] 384 True
132
+ │ │ └─MultiHeadAttention: 3-80 [32, 1024, 384] [32, 1024, 384] 589,952 True
133
+ │ │ └─RMSNorm: 3-81 [32, 1024, 384] [32, 1024, 384] 384 True
134
+ │ │ └─RMSNorm: 3-82 [32, 1024, 384] [32, 1024, 384] 384 True
135
+ │ │ └─SwiGLUMLP: 3-83 [32, 1024, 384] [32, 1024, 384] 884,736 True
136
+ │ │ └─RMSNorm: 3-84 [32, 1024, 384] [32, 1024, 384] 384 True
137
+ ├─RMSNorm: 1-4 [32, 1024, 384] [32, 1024, 384] 384 True
138
+ ├─Linear: 1-5 [32, 1024, 384] [32, 1024, 50258] 19,299,072 True
139
+ ============================================================================================================================================
140
+ Total params: 59,265,664
141
+ Trainable params: 59,265,664
142
+ Non-trainable params: 0
143
+ Total mult-adds (Units.GIGABYTES): 1.90
144
+ ============================================================================================================================================
145
+ Input size (MB): 0.26
146
+ Forward/backward pass size (MB): 34515.45
147
+ Params size (MB): 237.06
148
+ Estimated Total Size (MB): 34752.78
149
+ ============================================================================================================================================
150
+ ============================================================
151
+ 2026-04-08 13:51:28 - INFO - parrotllm.training - Starting evaluation...
152
+ 2026-04-08 13:51:34 - INFO - parrotllm.training - Initial val: loss=10.9068, ppl=54548.17
153
+ 2026-04-08 13:51:37 - INFO - parrotllm.training -
154
+ ============================================================
155
+ Starting training...
156
+ Steps per epoch (approx): 3
157
+ Max steps: 12
158
+ ============================================================
159
+ 2026-04-08 13:51:37 - INFO - parrotllm.training - LR schedule=wsd | decay_ratio=0.179 | z_loss_coeff=0.000275
160
+ 2026-04-08 13:51:58 - INFO - parrotllm.training - step 1 | epoch 0 | loss 10.9312 | lr 3.26e-04 | grad 6.6739
161
+ 2026-04-08 13:51:58 - INFO - parrotllm.training - Starting evaluation...
162
+ 2026-04-08 13:51:58 - INFO - parrotllm.training - ------------------------------------------------------------
163
+ 2026-04-08 13:52:01 - INFO - parrotllm.training - Train: loss=10.9312, ppl=55895.18
164
+ 2026-04-08 13:52:01 - INFO - parrotllm.training - Val: loss=10.5843, ppl=39509.51
165
+ 2026-04-08 13:52:01 - INFO - parrotllm.training - ** New best validation loss! **
166
+ 2026-04-08 13:52:02 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_10p5843_epoch_0000_step_0000001.pt
167
+ 2026-04-08 13:52:03 - INFO - parrotllm.training - ------------------------------------------------------------
168
+ 2026-04-08 13:52:04 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0000_step_0000001.pt
169
+ 2026-04-08 13:52:05 - INFO - parrotllm.training - step 2 | epoch 0 | loss 10.6100 | lr 6.52e-04 | grad 15.2774
170
+ 2026-04-08 13:52:05 - INFO - parrotllm.training - Starting evaluation...
171
+ 2026-04-08 13:52:05 - INFO - parrotllm.training - ------------------------------------------------------------
172
+ 2026-04-08 13:52:07 - INFO - parrotllm.training - Train: loss=10.6100, ppl=40540.20
173
+ 2026-04-08 13:52:07 - INFO - parrotllm.training - Val: loss=10.2400, ppl=28000.17
174
+ 2026-04-08 13:52:07 - INFO - parrotllm.training - ** New best validation loss! **
175
+ 2026-04-08 13:52:10 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_10p2400_epoch_0000_step_0000002.pt
176
+ 2026-04-08 13:52:11 - INFO - parrotllm.training - ------------------------------------------------------------
177
+ 2026-04-08 13:52:13 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0000_step_0000002.pt
178
+ 2026-04-08 13:52:13 - INFO - parrotllm.training - step 3 | epoch 0 | loss 10.2649 | lr 9.79e-04 | grad 1.7817
179
+ 2026-04-08 13:52:13 - INFO - parrotllm.training - Starting evaluation...
180
+ 2026-04-08 13:52:13 - INFO - parrotllm.training - ------------------------------------------------------------
181
+ 2026-04-08 13:52:16 - INFO - parrotllm.training - Epoch 1 complete:
182
+ 2026-04-08 13:52:16 - INFO - parrotllm.training - Train: loss=10.2649, ppl=28706.13
183
+ 2026-04-08 13:52:16 - INFO - parrotllm.training - Val: loss=9.9060, ppl=20049.88
184
+ 2026-04-08 13:52:16 - INFO - parrotllm.training - ** New best validation loss! **
185
+ 2026-04-08 13:52:17 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p9060_epoch_0001_step_0000003.pt
186
+ 2026-04-08 13:52:18 - INFO - parrotllm.training - ------------------------------------------------------------
187
+ 2026-04-08 13:52:19 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000003.pt
188
+ 2026-04-08 13:52:20 - INFO - parrotllm.training - step 4 | epoch 1 | loss 9.9116 | lr 1.30e-03 | grad 1.6496
189
+ 2026-04-08 13:52:20 - INFO - parrotllm.training - Starting evaluation...
190
+ 2026-04-08 13:52:20 - INFO - parrotllm.training - ------------------------------------------------------------
191
+ 2026-04-08 13:52:23 - INFO - parrotllm.training - Train: loss=9.9116, ppl=20162.53
192
+ 2026-04-08 13:52:23 - INFO - parrotllm.training - Val: loss=9.4788, ppl=13080.02
193
+ 2026-04-08 13:52:23 - INFO - parrotllm.training - ** New best validation loss! **
194
+ 2026-04-08 13:52:25 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p4788_epoch_0001_step_0000004.pt
195
+ 2026-04-08 13:52:26 - INFO - parrotllm.training - ------------------------------------------------------------
196
+ 2026-04-08 13:52:28 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000004.pt
197
+ 2026-04-08 13:52:29 - INFO - parrotllm.training - step 5 | epoch 1 | loss 9.4874 | lr 1.30e-03 | grad 1.5931
198
+ 2026-04-08 13:52:29 - INFO - parrotllm.training - Starting evaluation...
199
+ 2026-04-08 13:52:29 - INFO - parrotllm.training - ------------------------------------------------------------
200
+ 2026-04-08 13:52:32 - INFO - parrotllm.training - Train: loss=9.4874, ppl=13192.05
201
+ 2026-04-08 13:52:32 - INFO - parrotllm.training - Val: loss=9.0758, ppl=8741.00
202
+ 2026-04-08 13:52:32 - INFO - parrotllm.training - ** New best validation loss! **
203
+ 2026-04-08 13:52:33 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_9p0758_epoch_0001_step_0000005.pt
204
+ 2026-04-08 13:52:34 - INFO - parrotllm.training - ------------------------------------------------------------
205
+ 2026-04-08 13:52:36 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0001_step_0000005.pt
206
+ 2026-04-08 13:52:37 - INFO - parrotllm.training - step 6 | epoch 1 | loss 9.0752 | lr 1.30e-03 | grad 1.5725
207
+ 2026-04-08 13:52:37 - INFO - parrotllm.training - Starting evaluation...
208
+ 2026-04-08 13:52:37 - INFO - parrotllm.training - ------------------------------------------------------------
209
+ 2026-04-08 13:52:40 - INFO - parrotllm.training - Epoch 2 complete:
210
+ 2026-04-08 13:52:40 - INFO - parrotllm.training - Train: loss=9.0752, ppl=8735.76
211
+ 2026-04-08 13:52:40 - INFO - parrotllm.training - Val: loss=8.7135, ppl=6084.21
212
+ 2026-04-08 13:52:40 - INFO - parrotllm.training - ** New best validation loss! **
213
+ 2026-04-08 13:52:43 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p7135_epoch_0002_step_0000006.pt
214
+ 2026-04-08 13:52:43 - INFO - parrotllm.training - ------------------------------------------------------------
215
+ 2026-04-08 13:52:46 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000006.pt
216
+ 2026-04-08 13:52:47 - INFO - parrotllm.training - step 7 | epoch 2 | loss 8.7097 | lr 1.30e-03 | grad 1.5173
217
+ 2026-04-08 13:52:47 - INFO - parrotllm.training - Starting evaluation...
218
+ 2026-04-08 13:52:47 - INFO - parrotllm.training - ------------------------------------------------------------
219
+ 2026-04-08 13:52:49 - INFO - parrotllm.training - Train: loss=8.7097, ppl=6061.44
220
+ 2026-04-08 13:52:49 - INFO - parrotllm.training - Val: loss=8.4107, ppl=4494.89
221
+ 2026-04-08 13:52:49 - INFO - parrotllm.training - ** New best validation loss! **
222
+ 2026-04-08 13:52:51 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p4107_epoch_0002_step_0000007.pt
223
+ 2026-04-08 13:52:51 - INFO - parrotllm.training - ------------------------------------------------------------
224
+ 2026-04-08 13:52:52 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000007.pt
225
+ 2026-04-08 13:52:53 - INFO - parrotllm.training - step 8 | epoch 2 | loss 8.4315 | lr 1.30e-03 | grad 1.4582
226
+ 2026-04-08 13:52:53 - INFO - parrotllm.training - Starting evaluation...
227
+ 2026-04-08 13:52:53 - INFO - parrotllm.training - ------------------------------------------------------------
228
+ 2026-04-08 13:52:56 - INFO - parrotllm.training - Train: loss=8.4315, ppl=4589.18
229
+ 2026-04-08 13:52:56 - INFO - parrotllm.training - Val: loss=8.1591, ppl=3495.20
230
+ 2026-04-08 13:52:56 - INFO - parrotllm.training - ** New best validation loss! **
231
+ 2026-04-08 13:52:58 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_8p1591_epoch_0002_step_0000008.pt
232
+ 2026-04-08 13:52:59 - INFO - parrotllm.training - ------------------------------------------------------------
233
+ 2026-04-08 13:53:01 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0002_step_0000008.pt
234
+ 2026-04-08 13:53:02 - INFO - parrotllm.training - step 9 | epoch 2 | loss 8.1407 | lr 1.30e-03 | grad 1.3348
235
+ 2026-04-08 13:53:02 - INFO - parrotllm.training - Starting evaluation...
236
+ 2026-04-08 13:53:02 - INFO - parrotllm.training - ------------------------------------------------------------
237
+ 2026-04-08 13:53:05 - INFO - parrotllm.training - Epoch 3 complete:
238
+ 2026-04-08 13:53:05 - INFO - parrotllm.training - Train: loss=8.1407, ppl=3431.31
239
+ 2026-04-08 13:53:05 - INFO - parrotllm.training - Val: loss=7.9587, ppl=2860.23
240
+ 2026-04-08 13:53:05 - INFO - parrotllm.training - ** New best validation loss! **
241
+ 2026-04-08 13:53:06 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p9587_epoch_0003_step_0000009.pt
242
+ 2026-04-08 13:53:07 - INFO - parrotllm.training - ------------------------------------------------------------
243
+ 2026-04-08 13:53:10 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000009.pt
244
+ 2026-04-08 13:53:11 - INFO - parrotllm.training - step 10 | epoch 3 | loss 7.9377 | lr 1.30e-03 | grad 1.1490
245
+ 2026-04-08 13:53:11 - INFO - parrotllm.training - Starting evaluation...
246
+ 2026-04-08 13:53:11 - INFO - parrotllm.training - ------------------------------------------------------------
247
+ 2026-04-08 13:53:14 - INFO - parrotllm.training - Train: loss=7.9377, ppl=2800.99
248
+ 2026-04-08 13:53:14 - INFO - parrotllm.training - Val: loss=7.8117, ppl=2469.28
249
+ 2026-04-08 13:53:14 - INFO - parrotllm.training - ** New best validation loss! **
250
+ 2026-04-08 13:53:16 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p8117_epoch_0003_step_0000010.pt
251
+ 2026-04-08 13:53:17 - INFO - parrotllm.training - ------------------------------------------------------------
252
+ 2026-04-08 13:53:19 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000010.pt
253
+ 2026-04-08 13:53:20 - INFO - parrotllm.training - step 11 | epoch 3 | loss 7.7149 | lr 7.18e-04 | grad 0.8942
254
+ 2026-04-08 13:53:20 - INFO - parrotllm.training - Starting evaluation...
255
+ 2026-04-08 13:53:20 - INFO - parrotllm.training - ------------------------------------------------------------
256
+ 2026-04-08 13:53:22 - INFO - parrotllm.training - Train: loss=7.7149, ppl=2241.57
257
+ 2026-04-08 13:53:22 - INFO - parrotllm.training - Val: loss=7.7523, ppl=2326.87
258
+ 2026-04-08 13:53:22 - INFO - parrotllm.training - ** New best validation loss! **
259
+ 2026-04-08 13:53:24 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p7523_epoch_0003_step_0000011.pt
260
+ 2026-04-08 13:53:24 - INFO - parrotllm.training - ------------------------------------------------------------
261
+ 2026-04-08 13:53:25 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0003_step_0000011.pt
262
+ 2026-04-08 13:53:26 - INFO - parrotllm.training - step 12 | epoch 3 | loss 7.6396 | lr 1.30e-04 | grad 0.7553
263
+ 2026-04-08 13:53:26 - INFO - parrotllm.training - Starting evaluation...
264
+ 2026-04-08 13:53:26 - INFO - parrotllm.training - ------------------------------------------------------------
265
+ 2026-04-08 13:53:29 - INFO - parrotllm.training - Epoch 4 complete:
266
+ 2026-04-08 13:53:29 - INFO - parrotllm.training - Train: loss=7.6396, ppl=2078.98
267
+ 2026-04-08 13:53:29 - INFO - parrotllm.training - Val: loss=7.7414, ppl=2301.75
268
+ 2026-04-08 13:53:29 - INFO - parrotllm.training - ** New best validation loss! **
269
+ 2026-04-08 13:53:32 - INFO - parrotllm.training - Saved best checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/best_loss_7p7414_epoch_0004_step_0000012.pt
270
+ 2026-04-08 13:53:32 - INFO - parrotllm.training - ------------------------------------------------------------
271
+ 2026-04-08 13:53:35 - INFO - parrotllm.training - Saved checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0004_step_0000012.pt
272
+ 2026-04-08 13:53:37 - INFO - parrotllm.training - Saved final checkpoint: /home/user/code/ParrotLLM/runs/smoketest/exp_c/run_20260408_135126/checkpoints/last_epoch_0004_step_0000012.pt
273
+ 2026-04-08 13:53:37 - INFO - parrotllm.training -
274
+ Loss Curve:
275
+ Loss
276
+ ^
277
+ 10.93 |*
278
+ | *
279
+ | *
280
+ | *
281
+ | * *
282
+ | *
283
+ | * *
284
+ 7.64 | * * *
285
+ +--------------------------------------------------> step
286
+ 1 12
287
+ 2026-04-08 13:53:37 - INFO - parrotllm.training -
288
+ ============================================================
289
+ TRAINING COMPLETE
290
+ ============================================================
291
+ Epochs: 4
292
+ Total steps: 12
293
+ Total time: 0.03 hours
294
+ Best validation loss: 7.7414
295
+ Run directory: runs/smoketest/exp_c/run_20260408_135126
296
+ ============================================================
297
+ 2026-04-08 13:53:37 - INFO - parrotllm.training - Uploading completed run directory to Hugging Face: repo=ParrotLabs/Preprocessed
runs/smoketest/exp_c/run_20260408_135126/training_plots.pdf ADDED
Binary file (21.3 kB). View file