agent-sandbox enables easy management of isolated, stateful, singleton workloads, ideal for use cases like AI agent runtimes and reinforcement learning (RL).
Rl GitHub Repositories (2026) · Open Source Tools
Discover the best Rl open source repositories on GitHub. Explore trending tools, live star counts, source code, and developer projects updated daily.
Libra: Efficient Resource Management for Agentic RL Post-Training
The RL Bridge for LLM-based Agent Applications. Made Simple & Flexible.
A modular, primitive-first, python-first PyTorch library for Reinforcement Learning.
A native Mac App for LLM fine-tuning on Apple Silicon — fully on-device, fully open source.
Tensorlake is a serverless runtime for sandboxes and deploying background agentic applications
Python library for Reinforcement Learning.
Scalable, Composable Reinforcement Learning for Real-World Agents with Bedrock AgentCore.