OpenRLHF: Ray-based multi-model RLHF scheduling

What changed

Separated and scheduled RLHF model roles using Ray, vLLM and DeepSpeed, integrating large-model training and multiple alignment algorithms in an accessible framework.

Scientific scope

Separated and scheduled RLHF model roles using Ray, vLLM and DeepSpeed, integrating large-model training and multiple alignment algorithms in an accessible framework. Original six-person technical report includes the single-name author Xianyu; affiliation labels are not extra authors. Implements inherited PPO/DPO/KTO and systems components, not invention of those algorithms or guaranteed human alignment. Original2024features do not include every later repository capability.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Original six-person technical report includes the single-name author Xianyu; affiliation labels are not extra authors. Implements inherited PPO/DPO/KTO and systems components, not invention of those algorithms or guaranteed human alignment. Original2024features do not include every later repository capability.

Contributors

• Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.

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

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction