Titan AI LogoTitan AI

pysc2

8,209
1,172
Python

Project Description

StarCraft II Learning Environment

pysc2: StarCraft II Learning Environment

Project Title

pysc2 — Python-based StarCraft II Learning Environment for AI Research

Overview

PySC2 is a Python component of the StarCraft II Learning Environment (SC2LE) developed by DeepMind in collaboration with Blizzard Entertainment. It exposes StarCraft II's Machine Learning API as a Python RL Environment, providing an interface for RL agents to interact with StarCraft II, getting observations and sending actions. This project is unique in its focus on using the complex game of StarCraft II for Deep Reinforcement Learning (DeepRL) research.

Key Features

  • Exposes StarCraft II Machine Learning API as a Python RL Environment
  • Enables AI agents to interact with StarCraft II for RL research
  • Provides an interface for getting observations and sending actions within the game

Use Cases

  • Researchers using PySC2 for Deep Reinforcement Learning experiments
  • Academics studying complex decision-making processes in AI through the game's strategic elements
  • Game developers interested in AI behavior and machine learning integration within game environments

Advantages

  • Rich environment for AI research due to the complexity of StarCraft II
  • Collaboration between DeepMind and Blizzard ensures robust API support
  • Python-based, making it accessible to a wide range of developers and researchers

Limitations / Considerations

  • Requires the full version of StarCraft II game with the API included (version 3.16.1 and above)
  • Primarily tested on Linux, with limited testing on MacOS and Windows systems
  • May require additional setup and configuration depending on the system and game installation

Similar / Related Projects

  • StarCraft II Machine Learning API: The underlying API that PySC2 interfaces with, developed by Blizzard.
  • OpenAI Gym: A toolkit for developing and comparing reinforcement learning algorithms, with a different set of environments.
  • VizDoom: A Doom-based AI research platform, providing a different game environment for AI research.

Basic Information


📊 Project Information

🏷️ Project Topics

Topics: [, ", b, l, i, z, z, a, r, d, -, a, p, i, ", ,, , ", d, e, e, p, m, i, n, d, ", ,, , ", m, a, c, h, i, n, e, -, l, e, a, r, n, i, n, g, ", ,, , ", r, e, i, n, f, o, r, c, e, m, e, n, t, -, l, e, a, r, n, i, n, g, ", ,, , ", s, t, a, r, c, r, a, f, t, -, i, i, ", ,, , ", s, t, a, r, c, r, a, f, t, -, i, i, -, r, e, p, l, a, y, s, ", ]



This article is automatically generated by AI based on GitHub project information and README content analysis

Titan AI Explorehttps://www.titanaiexplore.com/projects/pysc2-98336690en-USTechnology

Project Information

Created on 7/25/2017
Updated on 11/15/2025