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: 2.5347
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: 32
- eval_batch_size: 32
- seed: 1234
- gradient_accumulation_steps: 16
- 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: 20
- num_epochs: 20.0
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 4.1842 | 0.5093 | 160 | 4.1144 |
| 3.8468 | 1.0159 | 320 | 3.8208 |
| 3.6684 | 1.5252 | 480 | 3.5943 |
| 3.4693 | 2.0318 | 640 | 3.3978 |
| 3.3095 | 2.5411 | 800 | 3.2516 |
| 3.1952 | 3.0477 | 960 | 3.1402 |
| 3.0920 | 3.5570 | 1120 | 3.0440 |
| 3.0295 | 4.0637 | 1280 | 2.9647 |
| 2.9459 | 4.5729 | 1440 | 2.8963 |
| 2.8952 | 5.0796 | 1600 | 2.8432 |
| 2.8762 | 5.5888 | 1760 | 2.8007 |
| 2.7562 | 6.0955 | 1920 | 2.7677 |
| 2.7841 | 6.6047 | 2080 | 2.7405 |
| 2.7474 | 7.1114 | 2240 | 2.7152 |
| 2.7327 | 7.6206 | 2400 | 2.6953 |
| 2.6648 | 8.1273 | 2560 | 2.6761 |
| 2.6745 | 8.6366 | 2720 | 2.6611 |
| 2.6793 | 9.1432 | 2880 | 2.6471 |
| 2.6427 | 9.6525 | 3040 | 2.6336 |
| 2.6231 | 10.1591 | 3200 | 2.6207 |
| 2.6001 | 10.6684 | 3360 | 2.6115 |
| 2.5883 | 11.1751 | 3520 | 2.6026 |
| 2.6036 | 11.6843 | 3680 | 2.5933 |
| 2.5517 | 12.1910 | 3840 | 2.5865 |
| 2.5718 | 12.7002 | 4000 | 2.5794 |
| 2.5626 | 13.2069 | 4160 | 2.5744 |
| 2.5513 | 13.7161 | 4320 | 2.5686 |
| 2.5427 | 14.2228 | 4480 | 2.5636 |
| 2.5487 | 14.7320 | 4640 | 2.5597 |
| 2.5285 | 15.2387 | 4800 | 2.5553 |
| 2.5360 | 15.7480 | 4960 | 2.5519 |
| 2.5122 | 16.2546 | 5120 | 2.5484 |
| 2.5207 | 16.7639 | 5280 | 2.5456 |
| 2.5247 | 17.2705 | 5440 | 2.5436 |
| 2.4879 | 17.7798 | 5600 | 2.5408 |
| 2.5145 | 18.2865 | 5760 | 2.5403 |
| 2.5035 | 18.7957 | 5920 | 2.5374 |
| 2.4867 | 19.3024 | 6080 | 2.5364 |
| 2.4974 | 19.8116 | 6240 | 2.5347 |
| 2.5248 | 20.0 | 6300 | 2.5348 |
Framework versions
- Transformers 5.13.0
- Pytorch 2.11.0+cu128
- Datasets 5.0.0
- Tokenizers 0.22.2