<!-- 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. -->
gpt2-2-dp-mod-aochild-cut
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 4.4109
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 |
---|---|---|---|
6.7147 | 0.27 | 500 | 5.6451 |
5.3609 | 0.54 | 1000 | 5.2108 |
5.0162 | 0.81 | 1500 | 4.9585 |
4.7627 | 1.08 | 2000 | 4.8126 |
4.5775 | 1.35 | 2500 | 4.7013 |
4.4856 | 1.62 | 3000 | 4.6034 |
4.4038 | 1.89 | 3500 | 4.5175 |
4.2252 | 2.16 | 4000 | 4.4775 |
4.1408 | 2.42 | 4500 | 4.4236 |
4.1136 | 2.69 | 5000 | 4.3721 |
4.0852 | 2.96 | 5500 | 4.3281 |
3.87 | 3.23 | 6000 | 4.3418 |
3.8651 | 3.5 | 6500 | 4.3062 |
3.8601 | 3.77 | 7000 | 4.2781 |
3.8091 | 4.04 | 7500 | 4.2785 |
3.5972 | 4.31 | 8000 | 4.2888 |
3.6301 | 4.58 | 8500 | 4.2678 |
3.6398 | 4.85 | 9000 | 4.2396 |
3.4906 | 5.12 | 9500 | 4.2803 |
3.3704 | 5.39 | 10000 | 4.2849 |
3.4008 | 5.66 | 10500 | 4.2718 |
3.4029 | 5.93 | 11000 | 4.2491 |
3.1804 | 6.2 | 11500 | 4.3116 |
3.1361 | 6.47 | 12000 | 4.3119 |
3.1532 | 6.73 | 12500 | 4.3067 |
3.1591 | 7.0 | 13000 | 4.3072 |
2.8974 | 7.27 | 13500 | 4.3563 |
2.9167 | 7.54 | 14000 | 4.3589 |
2.9248 | 7.81 | 14500 | 4.3580 |
2.8683 | 8.08 | 15000 | 4.3791 |
2.741 | 8.35 | 15500 | 4.3939 |
2.7503 | 8.62 | 16000 | 4.3968 |
2.7573 | 8.89 | 16500 | 4.3983 |
2.6961 | 9.16 | 17000 | 4.4075 |
2.6562 | 9.43 | 17500 | 4.4101 |
2.6653 | 9.7 | 18000 | 4.4107 |
2.667 | 9.97 | 18500 | 4.4109 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3