<!-- 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. -->
ss-100k-1
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 5.5173
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
- gradient_accumulation_steps: 8
- total_train_batch_size: 512
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 1000
- num_epochs: 18
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 9.4558 | 0.57 | 50 | 8.0579 |
| 7.7435 | 1.14 | 100 | 6.5844 |
| 6.4414 | 1.7 | 150 | 5.8475 |
| 5.9007 | 2.27 | 200 | 5.5338 |
| 5.5694 | 2.84 | 250 | 5.3088 |
| 5.3013 | 3.41 | 300 | 5.1569 |
| 5.0988 | 3.98 | 350 | 5.0276 |
| 4.8951 | 4.54 | 400 | 4.9282 |
| 4.7554 | 5.11 | 450 | 4.8577 |
| 4.5666 | 5.68 | 500 | 4.7978 |
| 4.4377 | 6.25 | 550 | 4.7625 |
| 4.3028 | 6.82 | 600 | 4.7126 |
| 4.1444 | 7.39 | 650 | 4.7261 |
| 4.0778 | 7.95 | 700 | 4.6753 |
| 3.8698 | 8.52 | 750 | 4.7045 |
| 3.839 | 9.09 | 800 | 4.7152 |
| 3.6258 | 9.66 | 850 | 4.7403 |
| 3.5713 | 10.23 | 900 | 4.7892 |
| 3.4446 | 10.79 | 950 | 4.7699 |
| 3.286 | 11.36 | 1000 | 4.8856 |
| 3.2562 | 11.93 | 1050 | 4.8461 |
| 2.9637 | 12.5 | 1100 | 4.9527 |
| 2.9602 | 13.07 | 1150 | 5.0062 |
| 2.6089 | 13.64 | 1200 | 5.0722 |
| 2.5281 | 14.2 | 1250 | 5.1845 |
| 2.2841 | 14.77 | 1300 | 5.2223 |
| 2.0771 | 15.34 | 1350 | 5.3392 |
| 1.9404 | 15.91 | 1400 | 5.3822 |
| 1.7145 | 16.48 | 1450 | 5.4698 |
| 1.6542 | 17.05 | 1500 | 5.4900 |
| 1.5286 | 17.61 | 1550 | 5.5173 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.12.0
- Tokenizers 0.13.3
