lucidrains/self-rewarding-lm-pytorch
A PyTorch training framework implementing MetaAI's Self-Rewarding Language Model with DPO and SPIN training approaches.
Not currently ranked — collecting fresh signals.
star history
This repository implements the training framework from MetaAI’s Self-Rewarding Language Model paper. It enables language models to iteratively improve themselves by serving as their own reward models during training. The framework incorporates Direct Preference Optimization (DPO) for preference-based training and also includes an implementation of the SPIN training method. Built on PyTorch with integration for transformer architectures.
Frequently asked
- What is lucidrains/self-rewarding-lm-pytorch?
- A PyTorch training framework implementing MetaAI's Self-Rewarding Language Model with DPO and SPIN training approaches.
- Is self-rewarding-lm-pytorch open source?
- Yes — lucidrains/self-rewarding-lm-pytorch is open source, released under the MIT license.
- What language is self-rewarding-lm-pytorch written in?
- lucidrains/self-rewarding-lm-pytorch is primarily written in Python.
- How popular is self-rewarding-lm-pytorch?
- lucidrains/self-rewarding-lm-pytorch has 1.4k stars on GitHub.
- Where can I find self-rewarding-lm-pytorch?
- lucidrains/self-rewarding-lm-pytorch is on GitHub at https://github.com/lucidrains/self-rewarding-lm-pytorch.