<!-- 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. -->
gpt2-2_left_out_gutenberg
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 3.9287
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 |
|---|---|---|---|
| 5.8917 | 0.26 | 500 | 5.0150 |
| 4.6559 | 0.53 | 1000 | 4.6338 |
| 4.3512 | 0.79 | 1500 | 4.4091 |
| 4.1461 | 1.06 | 2000 | 4.2691 |
| 3.9654 | 1.32 | 2500 | 4.1719 |
| 3.8972 | 1.59 | 3000 | 4.0869 |
| 3.8271 | 1.85 | 3500 | 4.0113 |
| 3.6889 | 2.12 | 4000 | 3.9762 |
| 3.586 | 2.38 | 4500 | 3.9376 |
| 3.5724 | 2.65 | 5000 | 3.8870 |
| 3.5435 | 2.91 | 5500 | 3.8480 |
| 3.3888 | 3.17 | 6000 | 3.8520 |
| 3.3327 | 3.44 | 6500 | 3.8282 |
| 3.3538 | 3.7 | 7000 | 3.8039 |
| 3.3427 | 3.97 | 7500 | 3.7743 |
| 3.1287 | 4.23 | 8000 | 3.8093 |
| 3.1293 | 4.5 | 8500 | 3.7959 |
| 3.1508 | 4.76 | 9000 | 3.7735 |
| 3.1169 | 5.03 | 9500 | 3.7815 |
| 2.8937 | 5.29 | 10000 | 3.8078 |
| 2.9281 | 5.56 | 10500 | 3.7999 |
| 2.9357 | 5.82 | 11000 | 3.7869 |
| 2.8489 | 6.08 | 11500 | 3.8165 |
| 2.6858 | 6.35 | 12000 | 3.8367 |
| 2.7074 | 6.61 | 12500 | 3.8300 |
| 2.7252 | 6.88 | 13000 | 3.8234 |
| 2.5862 | 7.14 | 13500 | 3.8661 |
| 2.4957 | 7.41 | 14000 | 3.8772 |
| 2.5091 | 7.67 | 14500 | 3.8791 |
| 2.5155 | 7.94 | 15000 | 3.8773 |
| 2.3794 | 8.2 | 15500 | 3.9064 |
| 2.349 | 8.47 | 16000 | 3.9130 |
| 2.3595 | 8.73 | 16500 | 3.9154 |
| 2.3579 | 8.99 | 17000 | 3.9160 |
| 2.2743 | 9.26 | 17500 | 3.9268 |
| 2.2753 | 9.52 | 18000 | 3.9287 |
| 2.2734 | 9.79 | 18500 | 3.9287 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3
