[ICML 2025] Roll the dice & look before you leap: Going beyond the creative limits of next-token prediction

siiRL: Shanghai Innovation Institute RL Framework for Advanced LLMs and Multi-Agent Systems

Tools and pipelines for automated LLM performance evaluation

The official implementation of Self-Play Fine-Tuning (SPIN)

f-PO: Generalizing Preference Optimization with f-divergence Minimization

A Large-Scale, High-Quality Math Dataset for Reinforcement Learning in Language Models

We introduce 'Thinking with Video', a new paradigm leveraging video generation for multimodal reasoning. Our VideoThinkBench shows that Sora-2 surpasses GPT5 by 10% on eyeballing puzzles and reache…

slime is an LLM post-training framework for RL Scaling.

[NeurIPS 2025] An official implementation of Flow-GRPO: Training Flow Matching Models via Online RL

A clean implementation based on AlphaZero for any game in any framework + tutorial + Othello/Gobang/TicTacToe/Connect4 and more

Litex is a simple formal language Learnable in 2 hours.

A curated list of resources on Reinforcement Learning with Verifiable Rewards (RLVR) and the reasoning capability boundary of Large Language Models (LLMs).

🔍 Explore curated resources on Reinforcement Learning with Verifiable Rewards (RLVR) and the reasoning limits of Large Language Models (LLMs).

Code repository for "RL Grokking Recipe: How RL Unlocks and Transfers New Algorithms in LLMs""

Offcial Code of EyeReal

[R]einforcement [L]earning from [M]odel-rewarded [T]hinking - code for the paper "Language Models That Think, Chat Better"

Calibrate the camera with ZhangZhengyou method (in both distortion case and no distortion case)

Python implementation of Zhang's camera calibration method

Implementation of Zhang, Z., "A Flexible New Technique for Camera Calibration" (2000).

Intro to Reinforcement Learning (强化学习纲要）

RLP: Reinforcement as a Pretraining Objective

The best ChatGPT that $100 can buy.

MMD, Hausdorff and Sinkhorn divergences scaled up to 1,000,000 samples.

implementation of Wasserstein Natural Policy Gradients and Wasserstein Natural Evolution Strategies

A library of reinforcement learning components and agents

Source code of our ICML 2025 paper "Flowing Datasets with Wasserstein over Wasserstein Gradient Flows"

Moved to https://github.com/HINTLab/LeFusion