ags-lc0

This project is Leela Chess Zero (Lc0), a neural network-based chess engine that uses deep reinforcement learning. It employs a Monte Carlo Tree Search guided by a neural network trained through self-

SKILL.md

GitHub

lc0

Author: gmh5225 Source: mcp-gamehacking/skills/ags-lc0

Description

This project is Leela Chess Zero (Lc0), a neural network-based chess engine that uses deep reinforcement learning. It employs a Monte Carlo Tree Search guided by a neural network trained through self-play, following the approach of DeepMind's AlphaZero. The C++ engine supports CUDA, cuDNN, and OpenCL backends for GPU-accelerated neural network evaluation. It is aimed at chess engine developers, AI researchers, and competitive chess players interested in neural network-based game AI.