<!-- 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-simple-wiki-mod
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 4.3273
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 |
---|---|---|---|
6.6721 | 0.29 | 500 | 5.6311 |
5.3162 | 0.59 | 1000 | 5.2012 |
4.9814 | 0.88 | 1500 | 4.9493 |
4.708 | 1.17 | 2000 | 4.8102 |
4.5523 | 1.47 | 2500 | 4.6918 |
4.4524 | 1.76 | 3000 | 4.5941 |
4.3303 | 2.06 | 3500 | 4.5209 |
4.1432 | 2.35 | 4000 | 4.4726 |
4.1182 | 2.64 | 4500 | 4.4154 |
4.0753 | 2.94 | 5000 | 4.3598 |
3.8614 | 3.23 | 5500 | 4.3514 |
3.8147 | 3.52 | 6000 | 4.3176 |
3.7996 | 3.82 | 6500 | 4.2839 |
3.6896 | 4.11 | 7000 | 4.2834 |
3.5307 | 4.4 | 7500 | 4.2783 |
3.5227 | 4.7 | 8000 | 4.2595 |
3.5108 | 4.99 | 8500 | 4.2484 |
3.3413 | 5.28 | 9000 | 4.2624 |
3.3338 | 5.58 | 9500 | 4.2605 |
3.3305 | 5.87 | 10000 | 4.2597 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3