gpt2_left_out_children_stories
This model is a fine-tuned version of gpt2 on the generator dataset. It achieves the following results on the evaluation set:
- Loss: 3.9100
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 |
|---|---|---|---|
| 5.9591 | 0.25 | 500 | 5.0479 |
| 4.7279 | 0.49 | 1000 | 4.6585 |
| 4.4359 | 0.74 | 1500 | 4.4292 |
| 4.2425 | 0.99 | 2000 | 4.2868 |
| 4.033 | 1.24 | 2500 | 4.1992 |
| 3.9641 | 1.48 | 3000 | 4.1132 |
| 3.8924 | 1.73 | 3500 | 4.0414 |
| 3.8217 | 1.98 | 4000 | 3.9709 |
| 3.637 | 2.22 | 4500 | 3.9537 |
| 3.6122 | 2.47 | 5000 | 3.9079 |
| 3.607 | 2.72 | 5500 | 3.8681 |
| 3.5795 | 2.96 | 6000 | 3.8275 |
| 3.3865 | 3.21 | 6500 | 3.8409 |
| 3.3863 | 3.46 | 7000 | 3.8184 |
| 3.3855 | 3.71 | 7500 | 3.7940 |
| 3.3706 | 3.95 | 8000 | 3.7687 |
| 3.1824 | 4.2 | 8500 | 3.8055 |
| 3.169 | 4.45 | 9000 | 3.7896 |
| 3.182 | 4.69 | 9500 | 3.7696 |
| 3.1913 | 4.94 | 10000 | 3.7502 |
| 2.9852 | 5.19 | 10500 | 3.7978 |
| 2.9624 | 5.43 | 11000 | 3.7932 |
| 2.9763 | 5.68 | 11500 | 3.7807 |
| 2.9833 | 5.93 | 12000 | 3.7653 |
| 2.7879 | 6.18 | 12500 | 3.8182 |
| 2.742 | 6.42 | 13000 | 3.8233 |
| 2.7644 | 6.67 | 13500 | 3.8171 |
| 2.7652 | 6.92 | 14000 | 3.8087 |
| 2.6062 | 7.16 | 14500 | 3.8516 |
| 2.5539 | 7.41 | 15000 | 3.8602 |
| 2.5546 | 7.66 | 15500 | 3.8614 |
| 2.5663 | 7.91 | 16000 | 3.8598 |
| 2.4621 | 8.15 | 16500 | 3.8868 |
| 2.3987 | 8.4 | 17000 | 3.8939 |
| 2.4081 | 8.65 | 17500 | 3.8966 |
| 2.4097 | 8.89 | 18000 | 3.8973 |
| 2.3573 | 9.14 | 18500 | 3.9061 |
| 2.3304 | 9.39 | 19000 | 3.9089 |
| 2.3266 | 9.63 | 19500 | 3.9098 |
| 2.3276 | 9.88 | 20000 | 3.9100 |
Framework versions
- Transformers 4.26.1
- Pytorch 1.11.0+cu113
- Datasets 2.13.0
- Tokenizers 0.13.3