pytorch/rl

Why this repo matters

• •Modular, primitive-first design enables flexible reinforcement learning research and production deployment using native PyTorch patterns.
• •Supports multi-agent and model-based RL with distributed computing capabilities for complex robotics and decision-making tasks.
• •Maintained by the official PyTorch team ensuring long-term compatibility and integration with the broader ecosystem.