Request to release Qwen2-VL-7B-GRPO checkpoint

[zandfj]

Hi, thanks for your great work on “R1-VL: Learning to Reason with Multimodal LMs via Step-wise GRPO”.

In the paper / supplementary material you mention the Qwen2-VL-7B-GRPO model.
Could you please release this checkpoint, or share whether there are plans to release it?

Thanks!

[jingyi0000]

Thanks for your interest! We did not keep this trained weight. You may reproduce the results using the provided data and the training code in R1-V.