List your company
ResearchFormerly WizWand

Video Question Answering on VIDAL-QA LLaVA-Hound in-domain (test)

68.5AccuracyMM1.5-Video
27.83638.39348.9559.507Sep 30, 2024
Updated 8mo ago

Evaluation results

MethodLinks
MM1.5-VideoModel Scale=7B, Protocol=SFT2024.0968.5
MM1.5-VideoModel Scale=3B, Protocol=SFT2024.0963.4
MM1.5-VideoModel Scale=1B, Protocol=SFT2024.0960.6
LLAVA-HOUND-SFTModel Scale=7B2024.0956.3
MM1.5-VideoModel Scale=7B, Protocol=Training-free2024.0948.7
MM1.5-VideoModel Scale=3B, Protocol=Training-free2024.0945.4
MM1.5-VideoModel Scale=1B, Protocol=Training-free2024.0942.6
Video-LLaVAModel Scale=7B2024.0934.3
Chat-UniViModel Scale=7B2024.0931.4
LLaMA-VIDModel Scale=7B2024.0930.6
Video-ChatGPTModel Scale=7B2024.0929.4