verl-project ORG

verl/HybridFlow: A Flexible and Efficient RL Post-Training Framework

Multimodal RL training framework for diffusion & omni models

A unified framework for building, running, and training general agents at scale.

A set of examples based on verl for end-to-end RL training recipes.

Open MinT training runtime on veRL

verl Zero-Mismatch Dense/MoE HuggingFace Rollout

Provide performance insight capabilities for RL frameworks.

VLA training framework of Volcano Engine

verl data curation

No description.

verl Ascend specific recipe

External training/inference backend plugin for various hardware platform

Ongoing research training transformer models at scale