List your company
ResearchFormerly WizWand

Video Captioning on MSVC

3.67CorrectnessGemini 1.5 Pro
1.76682.26092.7553.2491Jun 11, 2024
Updated 8mo ago

Evaluation results

MethodLinks
Gemini 1.5 ProModel Scale=Pro2024.063.673.52
Gemini 1.5 FlashModel Scale=Flash2024.063.463.24
VideoLLaMA2.1Model Scale=7B, #Frames=162024.062.872.81
VideoLLaMA2Model Scale=72B, #Frames=162024.062.712.67
GPT4-V2024.062.72.76
Reka-Core2024.062.612.73
VideoLLaMA2Model Scale=7B, #Frames=82024.062.572.61
VideoLLaMA2Model Scale=7B, #Frames=162024.062.532.59
VideoLLaMA2Model Scale=8x7B, #Frames=82024.062.532.56
LLaVA-NeXT-VideoModel Scale=7B, #Frames=322024.062.42.52
VideoChat2Model Scale=7B, #Frames=162024.062.012.1
Video-LLaVAModel Scale=7B, #Frames=82024.061.852.05
LLAMA-VIDModel Scale=7B, #Frames=1 fps2024.061.842.11