Is there a Thinking/RL variant of LLaVA-OneVision-2?

[pengzhansun]

Hi team, thanks for open-sourcing this great work!

I noticed that LLaVA-OneVision-1.5 has an RL-trained variant (LLaVA-OV-1.5-RL) that can generate explicit chain-of-thought reasoning traces. Are there any plans to release a similar Thinking or RL-based version for LLaVA-OneVision-2?

Thanks!