<!-- 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-16k
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 3.0048
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: 7
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 6.7244 | 0.29 | 500 | 5.6427 |
| 5.3807 | 0.59 | 1000 | 5.2015 |
| 5.0299 | 0.88 | 1500 | 4.9542 |
| 4.7547 | 1.18 | 2000 | 4.8112 |
| 4.5995 | 1.47 | 2500 | 4.6903 |
| 4.4925 | 1.77 | 3000 | 4.5871 |
| 4.3559 | 2.06 | 3500 | 4.5151 |
| 4.172 | 2.35 | 4000 | 4.4652 |
| 4.1472 | 2.65 | 4500 | 4.4101 |
| 4.1107 | 2.94 | 5000 | 4.3580 |
| 3.8923 | 3.24 | 5500 | 4.3639 |
| 3.8591 | 3.53 | 6000 | 4.3367 |
| 3.8499 | 3.83 | 6500 | 4.2978 |
| 3.7264 | 4.12 | 7000 | 4.3082 |
| 3.5775 | 4.41 | 7500 | 4.2995 |
| 3.5773 | 4.71 | 8000 | 4.2837 |
| 3.5606 | 5.0 | 8500 | 4.2698 |
| 3.3137 | 5.3 | 9000 | 4.2999 |
| 3.325 | 5.59 | 9500 | 4.2995 |
| 3.32 | 5.89 | 10000 | 4.2935 |
| 3.2194 | 6.18 | 10500 | 4.3090 |
| 3.1682 | 6.47 | 11000 | 4.3104 |
| 3.1689 | 6.77 | 11500 | 4.3110 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3
