Llama中文社区，实时汇总最新Llama学习资料，构建最好的中文Llama大模型开源生态，完全开源可商用

Dopamine is a research framework for fast prototyping of reinforcement learning algorithms.

An elegant PyTorch deep reinforcement learning library.

Agent Reinforcement Trainer: train multi-step agents for real-world tasks using GRPO. Give your agents on-the-job training. Reinforcement learning for Qwen2.5, Qwen3, Llama, and more!

An implementation of the AlphaZero algorithm for Gomoku (also called Gobang or Five in a Row)

ELF: a platform for game research with AlphaGoZero/AlphaZero reimplementation

A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.

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

MuZero

A training framework for Stable Baselines3 reinforcement learning agents, with hyperparameter optimization and pre-trained agents included.

Reinforcement Learning Coach by Intel AI Lab enables easy experimentation with state of the art Reinforcement Learning algorithms

A Survey of Reinforcement Learning for Large Reasoning Models

Scalable RL solution for advanced reasoning of language models

Implementation of papers in 100 lines of code.

[ICML 2017] TensorFlow code for Curiosity-driven Exploration for Deep Reinforcement Learning

Latest Advances on System-2 Reasoning

Implementation of all RL algorithms in a simpler way

A collection of 100+ pre-trained RL agents using Stable Baselines, training and hyperparameter optimization included.

Understanding R1-Zero-Like Training: A Critical Perspective

The open source post-building layer for agents. Our environment data and evals power agent post-training (RL, SFT) and monitoring.