<!-- 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-length-16k-rarity-all-4k-1p2k
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 3.1849
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 |
---|---|---|---|
6.7383 | 0.3 | 500 | 5.6385 |
5.3727 | 0.59 | 1000 | 5.1984 |
5.0316 | 0.89 | 1500 | 4.9474 |
4.7489 | 1.18 | 2000 | 4.7974 |
4.5919 | 1.48 | 2500 | 4.6733 |
4.481 | 1.77 | 3000 | 4.5743 |
4.3448 | 2.07 | 3500 | 4.5035 |
4.1586 | 2.36 | 4000 | 4.4505 |
4.131 | 2.66 | 4500 | 4.3894 |
4.0922 | 2.95 | 5000 | 4.3352 |
3.8662 | 3.25 | 5500 | 4.3390 |
3.8273 | 3.54 | 6000 | 4.3014 |
3.8116 | 3.84 | 6500 | 4.2720 |
3.6686 | 4.13 | 7000 | 4.2734 |
3.5444 | 4.43 | 7500 | 4.2662 |
3.5274 | 4.73 | 8000 | 4.2522 |
3.5039 | 5.02 | 8500 | 4.2497 |
3.3378 | 5.32 | 9000 | 4.2560 |
3.336 | 5.61 | 9500 | 4.2548 |
3.3376 | 5.91 | 10000 | 4.2538 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3