Classification on VGGSound V -> T
48Top-1 AccuracyPE_AV L
Evaluation Results
| Method | Links | |
|---|---|---|
| PE_AV LZero-shot=true, FPS=30, Joint Query Mode=Native joint embeddings2025.12 | 48 | |
| PE_AV LZero-shot=true, Frames=16, Joint Query Mode=Native joint embeddings2025.12 | 47.3 | |
| LanguageBindZero-shot=true2025.12 | 45.4 | |
| ImageBindZero-shot=true2025.12 | 40.4 |