OpenRLHF Files

An Easy-to-use, Scalable and High-performance RLHF Framework

This is an exact mirror of the OpenRLHF project, hosted at https://github.com/OpenRLHF/OpenRLHF. SourceForge is not affiliated with OpenRLHF.

The interactive file manager requires Javascript. Please enable it or use sftp or scp.
You may still browse the files here.

Name	Modified	Size	InfoDownloads / Week
Parent folder
README.md	2026-03-24	1.1 kB	0
Release v0.9.7 source code.tar.gz	2026-03-24	377.0 kB	0
Release v0.9.7 source code.zip	2026-03-24	427.9 kB	0
Totals: 3 Items		805.9 kB	0

What's Changed

Features

Support evaluation in async training by @xiaoxigua999 in 266fbc5
Support partial rollout in async mode by @xiaoxigua999 in 9bbfdc9
Support OAI server for multi-turn agentic RL by @xiaoxigua999 in ea4ac84

Improvements

Add length logs in eval by @xiaoxigua999 in 6639fd3
Bump vLLM by @xiaoxigua999 in ac28c5d

Fixes

Fix some bugs related to #1200 and improve math utilities by @xiaoxigua999 in 7c54c83

Source: README.md, updated 2026-03-24

Other Useful Business Software

Forever Free Full-Stack Observability | Grafana Cloud

Our generous forever free tier includes the full platform, including the AI Assistant, for 3 users with 10k metrics, 50GB logs, and 50GB traces.

Built on open standards like Prometheus and OpenTelemetry, Grafana Cloud includes Kubernetes Monitoring, Application Observability, Incident Response, plus the AI-powered Grafana Assistant. Get started with our generous free tier today.

Create free account

Try Google Cloud Risk-Free With $300 in Credit

No hidden charges. No surprise bills. Cancel anytime.

Use your credit across every product. Compute, storage, AI, analytics. When it runs out, 20+ products stay free. You only pay when you choose to.

Start Free

Forever Free Full-Stack Observability | Grafana Cloud

Our generous forever free tier includes the full platform, including the AI Assistant, for 3 users with 10k metrics, 50GB logs, and 50GB traces.

Create free account

Recommended Projects

TradeMaster
TradeMaster is an open-source platform for quantitative trading
Tensorforce
A TensorFlow library for applied reinforcement learning
Maja Machine Learning Framework
This project provides a framework for testing and comparing different machine learning algorithms (particularly reinforcement learning methods) in different scenarios. Its intended area of application is in research and education.
Gym
Toolkit for developing and comparing reinforcement learning algorithms
Gymnasium
An API standard for single-agent reinforcement learning environments

OpenRLHF Files

An Easy-to-use, Scalable and High-performance RLHF Framework

Get an email when there's a new version of OpenRLHF

What's Changed

Features

Improvements

Fixes