<!-- 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-cl-rarity-sampling
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 4.7272
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: 1
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 6.5968 | 0.06 | 500 | 5.8631 |
| 5.3522 | 0.12 | 1000 | 5.4526 |
| 5.0178 | 0.18 | 1500 | 5.2242 |
| 4.7929 | 0.24 | 2000 | 5.0785 |
| 4.6294 | 0.3 | 2500 | 4.9954 |
| 4.4985 | 0.36 | 3000 | 4.9155 |
| 4.3881 | 0.42 | 3500 | 4.8630 |
| 4.2829 | 0.49 | 4000 | 4.8285 |
| 4.1842 | 0.55 | 4500 | 4.7980 |
| 4.0945 | 0.61 | 5000 | 4.7664 |
| 4.0089 | 0.67 | 5500 | 4.7366 |
| 3.9271 | 0.73 | 6000 | 4.7190 |
| 3.8657 | 0.79 | 6500 | 4.6997 |
| 3.8177 | 0.85 | 7000 | 4.6877 |
| 3.7835 | 0.91 | 7500 | 4.6805 |
| 3.775 | 0.97 | 8000 | 4.6790 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3
