<!-- 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-top-for-aochildes-cbt-guten
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 2.8489
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: 8
- mixed_precision_training: Native AMP
Training results
| Training Loss | Epoch | Step | Validation Loss |
|---|---|---|---|
| 6.7161 | 0.3 | 500 | 5.6285 |
| 5.363 | 0.61 | 1000 | 5.1966 |
| 5.018 | 0.91 | 1500 | 4.9552 |
| 4.7431 | 1.22 | 2000 | 4.8212 |
| 4.6079 | 1.52 | 2500 | 4.6956 |
| 4.4964 | 1.83 | 3000 | 4.5935 |
| 4.3231 | 2.13 | 3500 | 4.5371 |
| 4.2022 | 2.44 | 4000 | 4.4762 |
| 4.1726 | 2.74 | 4500 | 4.4224 |
| 4.0936 | 3.05 | 5000 | 4.3953 |
| 3.8889 | 3.35 | 5500 | 4.3793 |
| 3.9038 | 3.66 | 6000 | 4.3473 |
| 3.9001 | 3.96 | 6500 | 4.3124 |
| 3.6516 | 4.27 | 7000 | 4.3375 |
| 3.6389 | 4.57 | 7500 | 4.3236 |
| 3.6389 | 4.88 | 8000 | 4.3032 |
| 3.4714 | 5.18 | 8500 | 4.3287 |
| 3.3752 | 5.48 | 9000 | 4.3283 |
| 3.3822 | 5.79 | 9500 | 4.3180 |
| 3.3011 | 6.09 | 10000 | 4.3389 |
| 3.148 | 6.4 | 10500 | 4.3481 |
| 3.1561 | 6.7 | 11000 | 4.3480 |
| 3.1487 | 7.01 | 11500 | 4.3483 |
| 3.0231 | 7.31 | 12000 | 4.3604 |
| 3.0291 | 7.62 | 12500 | 4.3613 |
| 3.0268 | 7.92 | 13000 | 4.3614 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3
