verl-project/verl-omni

[RFC] Support MiniMax H3 T2VA FlowGRPO on GPU and NPU about 6 hours ago

good first issue

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

[RFC] Support Real-Time Full-Duplex Omni-Modal Post-Training (MiniCPM-o) 2 days ago

help wanted

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

[RFC] Multi-Turn Visual Reflection SFT for BAGEL in verl-omni 5 days ago

good first issue help wanted agent

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

[RFC] Support Online Policy Distillation (OPD) for Diffusion Models 5 days ago

help wanted

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

Qwen3Omni DPO/GRPO/OPD 训练功能支持 10 days ago

good first issue

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

Model support of BOOGU-Image 11 days ago

help wanted

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

[RFC] Design and Implementation Plan for HunyuanImage 3.0 Support 14 days ago

help wanted

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

[RFC]: Qwen3-Omni RL Support on FSDP with vLLM-Omni rollout 16 days ago

good first issue

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

[RFC] Image-to-Image (I2I) Diffusion Model Integration for FlowGRPO 25 days ago

good first issue

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

[RFC] Formal Omni-Model Support in verl-omni 28 days ago

good first issue

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

[RFC] v0.1 Release Tracker about 1 month ago

help wanted

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

MetaX MACA support path for VeRL-Omni about 2 months ago

good first issue

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

[RFC] SD3.5-Medium Flow-GRPO OCR as the standard fast convergence test about 2 months ago

enhancement good first issue

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

[misc] chore: open-source readiness checklist 3 months ago

good first issue help wanted

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

[misc] chore: architectural risks from upstream coupling and narrow abstractions 3 months ago

enhancement help wanted

verl-project/verl-omni

465

Multimodal RL training framework for diffusion & omni models

Python

#diffusion-models#flow-matching#grpo#multimodal#qwen#reinforcement-learning#rlhf#vllm

Open Issues Need Help