metadata
license: mit
base_model: openai-community/gpt2-medium
tags:
- trl
- reward-trainer
- generated_from_trainer
metrics:
- accuracy
model-index:
- name: 2c2-reward-medium
results: []
2c2-reward-medium
This model is a fine-tuned version of openai-community/gpt2-medium on an unknown dataset. It achieves the following results on the evaluation set:
- Loss: 0.4456
- Accuracy: 0.7936
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: 1.41e-05
- train_batch_size: 256
- eval_batch_size: 256
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: constant_with_warmup
- lr_scheduler_warmup_steps: 20
- num_epochs: 3.0
Training results
| Training Loss | Epoch | Step | Validation Loss | Accuracy |
|---|---|---|---|---|
| 1.1138 | 0.2151 | 20 | 1.0849 | 0.5157 |
| 0.704 | 0.4301 | 40 | 0.7146 | 0.5622 |
| 0.5512 | 0.6452 | 60 | 0.5744 | 0.7084 |
| 0.5099 | 0.8602 | 80 | 0.5031 | 0.7478 |
| 0.5035 | 1.0753 | 100 | 0.4814 | 0.7639 |
| 0.4409 | 1.2903 | 120 | 0.4687 | 0.7671 |
| 0.4318 | 1.5054 | 140 | 0.4550 | 0.7831 |
| 0.4379 | 1.7204 | 160 | 0.4463 | 0.7871 |
| 0.4402 | 1.9355 | 180 | 0.4366 | 0.7863 |
| 0.3794 | 2.1505 | 200 | 0.4592 | 0.7936 |
| 0.3813 | 2.3656 | 220 | 0.4481 | 0.7920 |
| 0.3634 | 2.5806 | 240 | 0.4384 | 0.8016 |
| 0.3867 | 2.7957 | 260 | 0.4456 | 0.7936 |
Framework versions
- Transformers 4.41.2
- Pytorch 2.1.2
- Datasets 2.19.2
- Tokenizers 0.19.1