Mistral-v0.2-orpo

👁 image/jpeg

Mistral-v0.2-orpo is a fine-tuned version of the new Mistral-7B-v0.2 on argilla/distilabel-capybara-dpo-7k-binarized preference dataset using Odds Ratio Preference Optimization (ORPO). The model has been trained for 1 epoch. It took almost 8 hours on A100 GPU.

💥 LazyORPO

This model has been trained using LazyORPO. A colab notebook that makes the training process much easier. Based on ORPO paper

👁 image/png

🎭 What is ORPO?

Odds Ratio Preference Optimization (ORPO) proposes a new method to train LLMs by combining SFT and Alignment into a new objective (loss function), achieving state of the art results. Some highlights of this techniques are:

• 🧠 Reference model-free → memory friendly
• 🔄 Replaces SFT+DPO/PPO with 1 single method (ORPO)
• 🏆 ORPO Outperforms SFT, SFT+DPO on PHI-2, Llama 2, and Mistral
• 📊 Mistral ORPO achieves 12.20% on AlpacaEval2.0, 66.19% on IFEval, and 7.32 on MT-Bench out Hugging Face Zephyr Beta

💻 Usage

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer

torch.set_default_device("cuda")

model = AutoModelForCausalLM.from_pretrained("abideen/Mistral-v0.2-orpo", torch_dtype="auto", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("abideen/Mistral-v0.2-orpo", trust_remote_code=True)

inputs = tokenizer('''
 """
 Write a detailed analogy between mathematics and a lighthouse.
 """''', return_tensors="pt", return_attention_mask=False)

outputs = model.generate(**inputs, max_length=200)
text = tokenizer.batch_decode(outputs)[0]
print(text)

🏆 Evaluation

COMING SOON