llama2-7b-dpo-full-sft-wo-live_qa
This model is a fine-tuned version of Minbyul/llama2-7b-wo-live_qa-sft on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:
- Loss: 0.4649
- Rewards/chosen: -0.1737
- Rewards/rejected: -0.6052
- Rewards/accuracies: 0.9167
- Rewards/margins: 0.4315
- Logps/rejected: -682.3730
- Logps/chosen: -361.0932
- Logits/rejected: -0.6401
- Logits/chosen: -0.8344
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 | Validation Loss | Rewards/chosen | Rewards/rejected | Rewards/accuracies | Rewards/margins | Logps/rejected | Logps/chosen | Logits/rejected | Logits/chosen |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 0.3165 | 0.74 | 100 | 0.5069 | -0.1530 | -0.4700 | 0.875 | 0.3170 | -668.8459 | -359.0135 | -0.6288 | -0.8302 |
Framework versions
- Transformers 4.39.0.dev0
- Pytorch 2.1.2
- Datasets 2.14.6
- Tokenizers 0.15.2
- Downloads last month
- 8