A RLHF Framework Enhances OpenRLHF.

Recipes to train the self-rewarding reasoning LLMs.

On Memorization of Large Language Models in Logical Reasoning

A unified suite for generating elite reasoning problems and training high-performance LLMs, including pioneering attention-free architectures

A flexible and efficient training framework for large-scale alignment tasks

[ICML 2025 Oral] CodeI/O: Condensing Reasoning Patterns via Code Input-Output Prediction

[NeurIPS'25] Official codebase for "SWE-RL: Advancing LLM Reasoning via Reinforcement Learning on Open Software Evolution"

Ling is a MoE LLM provided and open-sourced by InclusionAI.

Extend OpenRLHF to support LMM RL training for reproduction of DeepSeek-R1 on multimodal tasks.

✨ A synthetic dataset generation framework that produces diverse coding questions and verifiable solutions - all in one framwork

Latest Advances on System-2 Reasoning

A bidirectional pipeline parallelism algorithm for computation-communication overlap in DeepSeek V3/R1 training.

Lightning-Fast RL for LLM Reasoning and Agents. Made Simple & Flexible.

SIFT: Grounding LLM Reasoning in Contexts via Stickers

AtCoder Library

EasyR1: An Efficient, Scalable, Multi-Modality RL Training Framework based on veRL

FlashMLA: Efficient Multi-head Latent Attention Kernels

minimal-cost for training 0.5B R1-Zero

RL algorithm: Advantage induced policy alignment

Reproduce R1 Zero on Logic Puzzle

ReST-MCTS*: LLM Self-Training via Process Reward Guided Tree Search (NeurIPS 2024)

Democratizing Reinforcement Learning for LLMs

Official Repo for Open-Reasoner-Zero

Application Kernel for Containers

The LLVM Project is a collection of modular and reusable compiler and toolchain technologies.

Slurm: A Highly Scalable Workload Manager

An open-source Java library for Constraint Programming