AIPython★ 612⑂ 16 forksApache-2.0
XYZ-AI-Lab/axrl
AxisRL: RL post-training
AxisRL is a framework for post‑training of agentic reinforcement learning, using SGLang for rollouts, Megatron for training, and adapted to real‑world agent workflows.
// KEY FEATURES
- Fast trajectory generation thanks to integration with SGLang rollout engine
- Scalable training of large models based on Megatron with parallelism support
- Adaptation to real‑world agent processes via plug‑in API and ready‑made connectors
- Flexible configuration of policies, reward functions, and training loops via YAML configs
New repositories every 30 minutes
REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.
Join on Telegram