<!-- 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-cocnat-aochildes-mod-sub-length-10k
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 4.3425
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 |
|---|---|---|---|
| 6.6933 | 0.29 | 500 | 5.6341 |
| 5.3469 | 0.59 | 1000 | 5.1996 |
| 4.9864 | 0.88 | 1500 | 4.9580 |
| 4.7189 | 1.18 | 2000 | 4.8083 |
| 4.5609 | 1.47 | 2500 | 4.6850 |
| 4.4523 | 1.77 | 3000 | 4.5821 |
| 4.317 | 2.06 | 3500 | 4.5146 |
| 4.1329 | 2.35 | 4000 | 4.4652 |
| 4.1086 | 2.65 | 4500 | 4.4071 |
| 4.0635 | 2.94 | 5000 | 4.3601 |
| 3.8482 | 3.24 | 5500 | 4.3553 |
| 3.8055 | 3.53 | 6000 | 4.3282 |
| 3.7859 | 3.83 | 6500 | 4.2926 |
| 3.6619 | 4.12 | 7000 | 4.2970 |
| 3.5196 | 4.41 | 7500 | 4.2933 |
| 3.5139 | 4.71 | 8000 | 4.2857 |
| 3.4905 | 5.0 | 8500 | 4.2710 |
| 3.3203 | 5.3 | 9000 | 4.2871 |
| 3.322 | 5.59 | 9500 | 4.2867 |
| 3.3172 | 5.89 | 10000 | 4.2863 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3
