verl-FL Documentation# verl-FL User Guide Overview Key Features Getting Started Requirements Docker Installation (Recommended) pip Installation Quick Installation Script Training Backends Rollout Backends Installation Custom Environment Setup AMD ROCm Support Ascend NPU Support RL Algorithms PPO (Proximal Policy Optimization) GRPO (Group Relative Policy Optimization) DAPO (Decoupled Alignment Policy Optimization) Other Algorithms Platform Abstraction Supported Platforms Adding a New Accelerator Dataset Format