Hi ,I noticed your data processing scripts includes the VLM3R dataset. Just curious,how does the performance compare between VLM-3R dataset and SPAR + LLaVA-Hound on VSI-bench? Also, would you mind sharing the processed training JSON for VLM3R? Thanks!
Hi ,I noticed your data processing scripts includes the VLM3R dataset. Just curious,how does the performance compare between VLM-3R dataset and SPAR + LLaVA-Hound on VSI-bench? Also, would you mind sharing the processed training JSON for VLM3R? Thanks!