selfbiorag-7b-dpo-full-sft-wo-kqa_golden
This model is a fine-tuned version of Minbyul/selfbiorag-7b-wo-kqa_golden-sft on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:
- Loss: 0.2401
- Rewards/chosen: -1.0928
- Rewards/rejected: -13.1704
- Rewards/accuracies: 0.8942
- Rewards/margins: 12.0777
- Logps/rejected: -2031.5652
- Logps/chosen: -567.3484
- Logits/rejected: -0.2100
- Logits/chosen: -0.3532
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: 5e-06
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- distributed_type: multi-GPU
- num_devices: 4
- gradient_accumulation_steps: 2
- total_train_batch_size: 64
- total_eval_batch_size: 32
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_ratio: 0.1
- num_epochs: 1
Training results
| Training Loss | Epoch | Step | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.249 | 0.31 | 100 | 0.3604 | -0.7724 | -7.8952 | 0.8942 | 7.1228 | -1504.0413 | -535.3107 | -0.2666 | -0.2359 |
| 0.1374 | 0.62 | 200 | 0.2389 | -0.9231 | -8.0656 | 0.9038 | 7.1425 | -1521.0862 | -550.3824 | -0.1753 | -0.2822 |
| 0.0982 | 0.92 | 300 | 0.2413 | -1.0961 | -13.1849 | 0.8942 | 12.0888 | -2033.0142 | -567.6829 | -0.2111 | -0.3569 |
Framework versions
- Transformers 4.39.0.dev0
- Pytorch 2.1.2
- Datasets 2.14.6
- Tokenizers 0.15.2