<!-- 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-qed-rarity-no-cut
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 4.3275
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 |
|---|---|---|---|
| 6.7002 | 0.29 | 500 | 5.6309 |
| 5.3451 | 0.58 | 1000 | 5.2082 |
| 5.0021 | 0.88 | 1500 | 4.9592 |
| 4.7266 | 1.17 | 2000 | 4.8110 |
| 4.5737 | 1.46 | 2500 | 4.6859 |
| 4.4727 | 1.75 | 3000 | 4.5796 |
| 4.3511 | 2.04 | 3500 | 4.5066 |
| 4.1544 | 2.34 | 4000 | 4.4568 |
| 4.1252 | 2.63 | 4500 | 4.3988 |
| 4.083 | 2.92 | 5000 | 4.3471 |
| 3.8825 | 3.21 | 5500 | 4.3454 |
| 3.8226 | 3.5 | 6000 | 4.3139 |
| 3.8118 | 3.8 | 6500 | 4.2766 |
| 3.7159 | 4.09 | 7000 | 4.2763 |
| 3.5383 | 4.38 | 7500 | 4.2702 |
| 3.5395 | 4.67 | 8000 | 4.2556 |
| 3.5257 | 4.96 | 8500 | 4.2454 |
| 3.3727 | 5.26 | 9000 | 4.2570 |
| 3.3469 | 5.55 | 9500 | 4.2567 |
| 3.3465 | 5.84 | 10000 | 4.2550 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3
