gpt2_small_AR_superbpe_8192_parallel10_42
This model was trained from scratch on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 4.1049
- Accuracy: 0.2909
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: 5e-05
- train_batch_size: 128
- eval_batch_size: 128
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- lr_scheduler_warmup_ratio: 0.1
- lr_scheduler_warmup_steps: 1000
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|
| 6.6876 | 1.0 | 1584 | 5.1443 | 0.2001 |
| 4.8561 | 2.0 | 3168 | 4.6374 | 0.2386 |
| 4.6287 | 3.0 | 4752 | 4.4302 | 0.2567 |
| 4.4002 | 4.0 | 6336 | 4.3189 | 0.2680 |
| 4.3228 | 5.0 | 7920 | 4.2463 | 0.2756 |
| 4.2139 | 6.0 | 9504 | 4.1945 | 0.2799 |
| 4.1461 | 7.0 | 11088 | 4.1579 | 0.2843 |
| 4.105 | 8.0 | 12672 | 4.1299 | 0.2874 |
| 4.0626 | 9.0 | 14256 | 4.1115 | 0.2895 |
| 4.0408 | 10.0 | 15840 | 4.1049 | 0.2909 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.8.0+cu128
- Datasets 4.1.1
- Tokenizers 0.19.1
