dist-classicrl 0.1.17

A high-performance Python package for distributed classical reinforcement learning algorithms with support for single-threaded, parallel, and MPI-distributed Q-Learning training.

A Python package for distributed classical reinforcement learning algorithms.

🎯 Project Description

dist_classicrl provides high-performance, scalable implementations of classic reinforcement learning algorithms with support for distributed training. The library focuses on Q-Learning with multiple execution modes: single-threaded, parallel (multiprocessing), and distributed (MPI) training.

🚀 Key Features

• 🚀 Multiple Execution Modes: Single-threaded, parallel, and MPI-distributed training

• ⚡ High Performance: Optimized implementations with vectorized operations and performance benchmarking

• 🎮 Multi-Agent Support: Built-in support for multi-agent environments

• 🔧 Flexible Architecture: Abstract base classes for easy extension and custom environments

• 🌐 Standards Compliant: Compatible with Gymnasium and PettingZoo environments (coming soon)

📚 Documentation

Essential Guides:

• 📦 Installation Guide - Setup instructions and requirements

• 📊 Performance Benchmarks - Benchmarking results and analysis

• 📚 Tutorials - Step-by-step guides for using the library

• 🤝 Contributing Guidelines - Development standards and contribution process

• 📄 License - License terms and usage rights

• 👥 Authors - Project contributors and maintainers

• 📜 Changelog - Project history and version changes

• 📜 Code of Conduct - Guidelines for participation and conduct

—