<!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. -->
gpt2-concat-aochildes-rarity-no-cut
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 4.3361
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: 6
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 6.6968 | 0.29 | 500 | 5.6436 |
| 5.341 | 0.59 | 1000 | 5.1917 |
| 4.994 | 0.88 | 1500 | 4.9570 |
| 4.7152 | 1.17 | 2000 | 4.8034 |
| 4.5612 | 1.47 | 2500 | 4.6851 |
| 4.4489 | 1.76 | 3000 | 4.5769 |
| 4.3245 | 2.05 | 3500 | 4.5082 |
| 4.1319 | 2.34 | 4000 | 4.4605 |
| 4.1053 | 2.64 | 4500 | 4.4020 |
| 4.0601 | 2.93 | 5000 | 4.3484 |
| 3.8535 | 3.22 | 5500 | 4.3509 |
| 3.8011 | 3.52 | 6000 | 4.3227 |
| 3.7818 | 3.81 | 6500 | 4.2900 |
| 3.6838 | 4.1 | 7000 | 4.2906 |
| 3.5185 | 4.4 | 7500 | 4.2858 |
| 3.5132 | 4.69 | 8000 | 4.2735 |
| 3.4972 | 4.98 | 8500 | 4.2611 |
| 3.3354 | 5.28 | 9000 | 4.2771 |
| 3.3221 | 5.57 | 9500 | 4.2766 |
| 3.3206 | 5.86 | 10000 | 4.2758 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3
