run2-babylm-original-1234-gpt2-small
This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 2.5255
Model description
More information needed
Intended uses & limitations
More information needed
Training and evaluation data
More information needed
Training procedure
Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 0.0002
- train_batch_size: 64
- eval_batch_size: 64
- seed: 1234
- gradient_accumulation_steps: 8
- total_train_batch_size: 512
- optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: cosine_with_min_lr
- lr_scheduler_warmup_steps: 100
- num_epochs: 20.0
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 4.2662 | 0.5091 | 160 | 4.1882 |
| 3.8608 | 1.0159 | 320 | 3.8359 |
| 3.6871 | 1.5251 | 480 | 3.6121 |
| 3.4753 | 2.0318 | 640 | 3.4000 |
| 3.3026 | 2.5410 | 800 | 3.2452 |
| 3.1835 | 3.0477 | 960 | 3.1260 |
| 3.0762 | 3.5569 | 1120 | 3.0267 |
| 3.0104 | 4.0636 | 1280 | 2.9446 |
| 2.9278 | 4.5728 | 1440 | 2.8780 |
| 2.8795 | 5.0796 | 1600 | 2.8293 |
| 2.8627 | 5.5887 | 1760 | 2.7873 |
| 2.7431 | 6.0955 | 1920 | 2.7551 |
| 2.7710 | 6.6046 | 2080 | 2.7263 |
| 2.7350 | 7.1114 | 2240 | 2.7031 |
| 2.7203 | 7.6205 | 2400 | 2.6839 |
| 2.6526 | 8.1273 | 2560 | 2.6649 |
| 2.6625 | 8.6364 | 2720 | 2.6494 |
| 2.6682 | 9.1432 | 2880 | 2.6354 |
| 2.6314 | 9.6523 | 3040 | 2.6232 |
| 2.6121 | 10.1591 | 3200 | 2.6107 |
| 2.5892 | 10.6683 | 3360 | 2.6014 |
| 2.5774 | 11.1750 | 3520 | 2.5928 |
| 2.5932 | 11.6842 | 3680 | 2.5837 |
| 2.5412 | 12.1909 | 3840 | 2.5766 |
| 2.5610 | 12.7001 | 4000 | 2.5696 |
| 2.5516 | 13.2068 | 4160 | 2.5640 |
| 2.5415 | 13.7160 | 4320 | 2.5591 |
| 2.5321 | 14.2228 | 4480 | 2.5537 |
| 2.5382 | 14.7319 | 4640 | 2.5495 |
| 2.5187 | 15.2387 | 4800 | 2.5458 |
| 2.5262 | 15.7478 | 4960 | 2.5422 |
| 2.5024 | 16.2546 | 5120 | 2.5396 |
| 2.5102 | 16.7637 | 5280 | 2.5363 |
| 2.5146 | 17.2705 | 5440 | 2.5347 |
| 2.4779 | 17.7796 | 5600 | 2.5315 |
| 2.5041 | 18.2864 | 5760 | 2.5308 |
| 2.4933 | 18.7955 | 5920 | 2.5280 |
| 2.4762 | 19.3023 | 6080 | 2.5271 |
| 2.4870 | 19.8115 | 6240 | 2.5255 |
| 2.5148 | 20.0 | 6300 | 2.5259 |
Framework versions
- Transformers 5.13.0
- Pytorch 2.11.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2