Minbyul's picture
End of training
0b3f71d verified
metadata
license: apache-2.0
base_model: Minbyul/biomistral-7b-wo-medication_qa-sft
tags:
  - alignment-handbook
  - trl
  - dpo
  - generated_from_trainer
  - trl
  - dpo
  - generated_from_trainer
datasets:
  - HuggingFaceH4/ultrafeedback_binarized
model-index:
  - name: biomistral-7b-dpo-full-sft-wo-medication_qa
    results: []

biomistral-7b-dpo-full-sft-wo-medication_qa

This model is a fine-tuned version of Minbyul/biomistral-7b-wo-medication_qa-sft on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4656
  • Rewards/chosen: -0.5066
  • Rewards/rejected: -1.3382
  • Rewards/accuracies: 0.75
  • Rewards/margins: 0.8316
  • Logps/rejected: -693.0344
  • Logps/chosen: -456.7123
  • Logits/rejected: -4.0821
  • Logits/chosen: -4.1231

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
0.1232 0.83 100 -4.1247 -4.0843 -456.2566 -691.8618 0.4671 0.75 -0.5020 0.8245 -1.3265

Framework versions

  • Transformers 4.39.0.dev0
  • Pytorch 2.1.2
  • Datasets 2.14.6
  • Tokenizers 0.15.2