metadata

license: apache-2.0
base_model: mistralai/Mistral-7B-v0.1
tags:
  - alignment-handbook
  - trl
  - dpo
  - generated_from_trainer
  - trl
  - dpo
  - generated_from_trainer
datasets:
  - HuggingFaceH4/ultrafeedback_binarized
model-index:
  - name: mistral-7b-dpo-full-wo-medication_qa-ep3
    results: []

mistral-7b-dpo-full-wo-medication_qa-ep3

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:

Loss: 0.3800
Rewards/chosen: -2.4212
Rewards/rejected: -6.9305
Rewards/accuracies: 0.8602
Rewards/margins: 4.5093
Logps/rejected: -1705.3644
Logps/chosen: -808.9231
Logits/rejected: -2.9073
Logits/chosen: -3.0809

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
0.1684	0.6	100	-3.1281	-2.9646	-755.8733	-1595.0167	0.3332	0.8686	-1.8907	3.9363	-5.8271

Framework versions

Transformers 4.39.0.dev0
Pytorch 2.1.2
Datasets 2.14.6
Tokenizers 0.15.2