mini-AlphaStar demystifies the complex AlphaStar architecture by offering an accessible, open-source PyTorch framework for StarCraft II AI development. It features spatial feature encoders for screen and minimap inputs, non-spatial feature processing, LSTM memory modules, and separate policy and value networks for action selection and state evaluation. Using imitation learning to bootstrap and reinforcement learning with self-play for fine-tuning, it supports environment wrappers compatible with StarCraft II via pysc2, logging through TensorBoard, and configurable hyperparameters. Researchers and students can generate datasets from human gameplay, train models on custom scenarios, evaluate agent performance, and visualize learning curves. The modular codebase enables easy experimentation with network variants, training schedules, and multi-agent setups. Designed for education and prototyping rather than production deployment.

This repository provides an end-to-end reinforcement learning framework for StarCraft II gameplay research. The core agent uses Proximal Policy Optimization (PPO) to learn policy networks that interpret observation data from the PySC2 environment and output precise in-game actions. Developers can configure neural network layers, reward shaping, and training schedules to optimize performance. The system supports multiprocessing for efficient sample collection, logging utilities for monitoring training curves, and evaluation scripts for running trained policies against scripted or built-in AI opponents. The codebase is written in Python and leverages TensorFlow for model definition and optimization. Users can extend components such as custom reward functions, state preprocessing, or network architectures to suit specific research objectives.