gpt2_left_out_open_subtitles
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 4.3147
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: 0.0005
- train_batch_size: 64
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 10
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 6.353 | 0.35 | 500 | 5.4389 |
| 5.0517 | 0.71 | 1000 | 4.9981 |
| 4.6553 | 1.06 | 1500 | 4.7472 |
| 4.3892 | 1.41 | 2000 | 4.5957 |
| 4.258 | 1.77 | 2500 | 4.4664 |
| 4.0843 | 2.12 | 3000 | 4.3902 |
| 3.922 | 2.48 | 3500 | 4.3282 |
| 3.89 | 2.83 | 4000 | 4.2537 |
| 3.713 | 3.18 | 4500 | 4.2408 |
| 3.6155 | 3.54 | 5000 | 4.2007 |
| 3.6227 | 3.89 | 5500 | 4.1522 |
| 3.4072 | 4.24 | 6000 | 4.1829 |
| 3.3651 | 4.6 | 6500 | 4.1499 |
| 3.3798 | 4.95 | 7000 | 4.1168 |
| 3.1038 | 5.3 | 7500 | 4.1744 |
| 3.1121 | 5.66 | 8000 | 4.1606 |
| 3.1104 | 6.01 | 8500 | 4.1563 |
| 2.8101 | 6.36 | 9000 | 4.2017 |
| 2.8457 | 6.72 | 9500 | 4.1979 |
| 2.7947 | 7.07 | 10000 | 4.2241 |
| 2.5839 | 7.43 | 10500 | 4.2544 |
| 2.6024 | 7.78 | 11000 | 4.2558 |
| 2.5182 | 8.13 | 11500 | 4.2838 |
| 2.41 | 8.49 | 12000 | 4.2963 |
| 2.416 | 8.84 | 12500 | 4.3009 |
| 2.3609 | 9.19 | 13000 | 4.3117 |
| 2.3182 | 9.55 | 13500 | 4.3138 |
| 2.3187 | 9.9 | 14000 | 4.3147 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3