meditron-7b-dpo-full-wo-kqa_silver_wogold-ep3
This model is a fine-tuned version of epfl-llm/meditron-7b on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:
- Loss: 0.5793
- Rewards/chosen: -0.1323
- Rewards/rejected: -0.4764
- Rewards/accuracies: 0.7717
- Rewards/margins: 0.3440
- Logps/rejected: -1456.3621
- Logps/chosen: -834.8738
- Logits/rejected: -0.9041
- Logits/chosen: -0.7062
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-07
- 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 | Logits/chosen | Logits/rejected | Logps/chosen | Logps/rejected | Validation Loss | Rewards/accuracies | Rewards/chosen | Rewards/margins | Rewards/rejected |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.5615 | 0.61 | 100 | -0.6676 | -0.8939 | -826.0934 | -1433.1564 | 0.6219 | 0.7459 | -0.0445 | 0.1998 | -0.2443 |
Framework versions
- Transformers 4.39.0.dev0
- Pytorch 2.1.2
- Datasets 2.14.6
- Tokenizers 0.15.2