OpenDILab Decision AI Engine. The Most Comprehensive Reinforcement Learning Framework B.P.

[NeurIPS 2023 Spotlight] LightZero: A Unified Benchmark for Monte Carlo Tree Search in General Sequential Decision Scenarios (awesome MCTS)

An artificial intelligence platform for the StarCraft II with large-scale distributed training and grand-master agents.

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

The official implementation of Self-Play Preference Optimization (SPPO)

A Massively Parallel Large Scale Self-Play Framework

SPIRAL: Self-Play on Zero-Sum Games Incentivizes Reasoning via Multi-Agent Multi-Turn Reinforcement Learning

Search Self-Play: Pushing the Frontier of Agent Capability without Supervision

AlphaZero implementation for Othello, Connect-Four and Tic-Tac-Toe based on "Mastering the game of Go without human knowledge" and "Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm" by DeepMind.

A very fast implementation of AlphaZero, applied to games like Splendor, Santorini, The Little Prince, … Browser version available

Backgammon OpenAI Gym

The exact codes used by the team "liveinparis" at the kaggle football competition ranked 6th/1141

TD-Gammon implementation

[ICLR'26] MARSHAL: Incentivizing Multi-Agent Reasoning via Self-Play with Strategic LLMs

AI agents for the bavarian card game Schafkopf trained with reinforcement learning

This is the implementation of paper Model Free Episodic Control

Using self-play, MCTS, and a deep neural network to create a hearthstone ai player

Code base for Social Robot Tree Search (SoRTS).

Distributed JAX self-play training for pgx environments - PPO, league play, baseline evaluation.

A gym environment to train chatbots.