InternLM/xtuner

About RLHF need

Open

#93 opened on Sep 4, 2023

View on GitHub
 (3 comments) (1 reaction) (1 assignee)Python (424 forks)github user discovery
feature requestgood first issue

Repository metrics

Stars
 (5,148 stars)
PR merge metrics
 (Avg merge 5d) (93 merged PRs in 30d)

Description

需要实现几种对齐算法 1.PPO 这个没的说，比较传统和通用，但是训练的开销会大一点 2. RAFT LMFLOW社区有做 https://optimalscale.github.io/LMFlow/examples/raft.html 3.pangu-coder2 RRTF (Rank Responses to align Test&Teacher Feedback) 总结一下是说，他们是用了代码单元测试，然后把单元测试的结果作为标签合并Loss微调LLM https://arxiv.org/abs/2307.14936 image image RRTF华为他们这部分没有开源。RAFT是开源了，RRTF可以的话可以一起讨论一起实现一下。

Contributor guide

Research direction
Study the referenced algorithms (PPO, RAFT, RRTF) and find open source implementations. For RRTF, analyze the paper and discuss with the community. Consider starting with RAFT as it has an open source reference.
Tech stack
python
Domain
machine learning
Issue type
Feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
4
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
Over 1 week
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
Active
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
Clear
Prerequisites
Python
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
20