paper-pretrained-model-gpt2
This model is a fine-tuned version of gpt2 on the None dataset. It achieves the following results on the evaluation set:
- Loss: 0.6482
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: 8
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 64
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 20
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| No log | 0.98 | 23 | 9.4651 |
| No log | 1.97 | 46 | 8.5615 |
| No log | 2.99 | 70 | 7.4620 |
| No log | 3.98 | 93 | 6.4971 |
| No log | 4.96 | 116 | 5.7243 |
| No log | 5.99 | 140 | 5.1366 |
| No log | 6.97 | 163 | 4.6595 |
| No log | 8.0 | 187 | 4.2285 |
| No log | 8.98 | 210 | 3.8841 |
| No log | 9.97 | 233 | 3.5671 |
| No log | 10.99 | 257 | 3.2337 |
| No log | 11.98 | 280 | 2.9197 |
| No log | 12.96 | 303 | 2.6052 |
| No log | 13.99 | 327 | 2.2564 |
| No log | 14.97 | 350 | 1.9214 |
| No log | 16.0 | 374 | 1.5827 |
| No log | 16.98 | 397 | 1.2635 |
| No log | 17.97 | 420 | 0.9987 |
| No log | 18.99 | 444 | 0.7605 |
| No log | 19.68 | 460 | 0.6482 |
Framework versions
- Transformers 4.38.2
- Pytorch 2.1.0+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2