biomistral-7b-wo-kqa_golden-iter-sft-dpo-step2
This model is a fine-tuned version of Minbyul/biomistral-7b-wo-kqa_golden-iter-sft-step2 on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:
- Loss: 0.6929
- Rewards/chosen: -0.5264
- Rewards/rejected: -0.6277
- Rewards/accuracies: 0.5278
- Rewards/margins: 0.1013
- Logps/rejected: -171.5833
- Logps/chosen: -175.8510
- Logits/rejected: -1.9829
- Logits/chosen: -2.0462
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 |
---|---|---|---|---|---|---|---|---|---|---|---|
0.5324 | 0.59 | 100 | 0.6768 | -0.2142 | -0.2788 | 0.5347 | 0.0646 | -136.6950 | -144.6288 | -2.2411 | -2.3325 |
Framework versions
- Transformers 4.39.0.dev0
- Pytorch 2.1.2
- Datasets 2.14.6
- Tokenizers 0.15.2
- Downloads last month
- 8
Inference API (serverless) is not available, repository is disabled.
Model tree for Minbyul/biomistral-7b-wo-kqa_golden-iter-sft-dpo-step2
Base model
BioMistral/BioMistral-7B
Finetuned
this model