https://github.com/howardchiang2/fine_tune_llama3_with_orpo

通过ORPO的方法微调llama3
https://github.com/howardchiang2/fine_tune_llama3_with_orpo

llama3 llm orpo

Last synced: 19 days ago
JSON representation

通过ORPO的方法微调llama3

• Host: GitHub
• URL: https://github.com/howardchiang2/fine_tune_llama3_with_orpo
• Owner: howardchiang2
• Created: 2024-05-09T01:19:41.000Z (9 months ago)
• Default Branch: main
• Last Pushed: 2024-05-09T01:41:49.000Z (9 months ago)
• Last Synced: 2025-01-12T04:32:23.569Z (20 days ago)
• Topics: llama3, llm, orpo
• Language: Jupyter Notebook
• Homepage:
• Size: 33.2 KB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

        
# fine_tune_llama3_with_ORPO  

通过ORPO的方法微调llama3  

创新点：结合SFT和对齐学习的新尝试  

环境：最好使用colab L4（需要开PRO）

参考论文《ORPO: Monolithic Preference Optimization without Reference Model》 

reference to blog:https://mlabonne.github.io/blog/posts/2024-04-19_Fine_tune_Llama_3_with_ORPO.html