<!-- 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-second
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 4.4031
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 |
|---|---|---|---|
| 6.7063 | 0.29 | 500 | 5.6161 |
| 5.3409 | 0.58 | 1000 | 5.1879 |
| 4.9975 | 0.87 | 1500 | 4.9292 |
| 4.7248 | 1.16 | 2000 | 4.7819 |
| 4.5625 | 1.45 | 2500 | 4.6577 |
| 4.4518 | 1.74 | 3000 | 4.5536 |
| 4.3506 | 2.02 | 3500 | 4.4718 |
| 4.1444 | 2.31 | 4000 | 4.4324 |
| 4.1299 | 2.6 | 4500 | 4.3859 |
| 4.097 | 2.89 | 5000 | 4.3383 |
| 3.9322 | 3.18 | 5500 | 4.3372 |
| 3.8738 | 3.47 | 6000 | 4.3092 |
| 3.8743 | 3.76 | 6500 | 4.2795 |
| 3.8147 | 4.05 | 7000 | 4.2758 |
| 3.6152 | 4.34 | 7500 | 4.2857 |
| 3.6479 | 4.63 | 8000 | 4.2632 |
| 3.654 | 4.92 | 8500 | 4.2380 |
| 3.4411 | 5.21 | 9000 | 4.2846 |
| 3.398 | 5.49 | 9500 | 4.2785 |
| 3.4249 | 5.78 | 10000 | 4.2628 |
| 3.3498 | 6.07 | 10500 | 4.2910 |
| 3.1525 | 6.36 | 11000 | 4.3119 |
| 3.1727 | 6.65 | 11500 | 4.3057 |
| 3.1862 | 6.94 | 12000 | 4.2985 |
| 2.9723 | 7.23 | 12500 | 4.3475 |
| 2.9448 | 7.52 | 13000 | 4.3551 |
| 2.9617 | 7.81 | 13500 | 4.3526 |
| 2.8946 | 8.1 | 14000 | 4.3748 |
| 2.7783 | 8.39 | 14500 | 4.3866 |
| 2.7819 | 8.68 | 15000 | 4.3904 |
| 2.7913 | 8.96 | 15500 | 4.3905 |
| 2.7052 | 9.25 | 16000 | 4.4009 |
| 2.6969 | 9.54 | 16500 | 4.4029 |
| 2.7 | 9.83 | 17000 | 4.4031 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3
