First of all, thank you for your excellent work on VideoChat-R1 — it's truly impressive!
I would like to kindly ask if it would be possible to release the SFT data used in the evaluation section, specifically the SFT training data based on Charades-STA, GoT-10k, and NExT-GQA.
Having access to these datasets would be incredibly helpful for reproducing your results and conducting further research. I would deeply appreciate it if you could consider making them available.
Thank you very much in advance!