README.md

metadata

license: apache-2.0
base_model: mistralai/Mistral-7B-v0.1
tags:
  - alignment-handbook
  - trl
  - orpo
  - generated_from_trainer
  - trl
  - orpo
  - generated_from_trainer
datasets:
  - orpo-explorers/OHP-15k-Stratified-1
  - orpo-explorers/OHP-15k-Stratified-2
  - orpo-explorers/OHP-15k-Stratified-3
model-index:
  - name: mistral-ohp-1-2-3-3epoch-beta-0.2
    results: []

mistral-ohp-1-2-3-3epoch-beta-0.2

This model is a fine-tuned version of mistralai/Mistral-7B-v0.1 on the orpo-explorers/OHP-15k-Stratified-1, the orpo-explorers/OHP-15k-Stratified-2 and the orpo-explorers/OHP-15k-Stratified-3 datasets.

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-06
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 2
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• total_eval_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: inverse_sqrt
• num_epochs: 3

Training results

Framework versions

• Transformers 4.39.3
• Pytorch 2.1.2.post303
• Datasets 2.18.0
• Tokenizers 0.15.2