Video Captioning on MSVC
3.67CorrectnessGemini 1.5 Pro
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Gemini 1.5 ProModel Scale=Pro2024.06 | 3.67 | 3.52 | |
| Gemini 1.5 FlashModel Scale=Flash2024.06 | 3.46 | 3.24 | |
| VideoLLaMA2.1Model Scale=7B, #Frames=162024.06 | 2.87 | 2.81 | |
| VideoLLaMA2Model Scale=72B, #Frames=162024.06 | 2.71 | 2.67 | |
| GPT4-V2024.06 | 2.7 | 2.76 | |
| Reka-Core2024.06 | 2.61 | 2.73 | |
| VideoLLaMA2Model Scale=7B, #Frames=82024.06 | 2.57 | 2.61 | |
| VideoLLaMA2Model Scale=7B, #Frames=162024.06 | 2.53 | 2.59 | |
| VideoLLaMA2Model Scale=8x7B, #Frames=82024.06 | 2.53 | 2.56 | |
| LLaVA-NeXT-VideoModel Scale=7B, #Frames=322024.06 | 2.4 | 2.52 | |
| VideoChat2Model Scale=7B, #Frames=162024.06 | 2.01 | 2.1 | |
| Video-LLaVAModel Scale=7B, #Frames=82024.06 | 1.85 | 2.05 | |
| LLAMA-VIDModel Scale=7B, #Frames=1 fps2024.06 | 1.84 | 2.11 |