Physical Audiovisual CommonSense Reasoning on PACS (test)
0.632AccuracyAudioCLIP w/ DCL
Evaluation Results
| Method | Links | |
|---|---|---|
| AudioCLIP w/ DCLZero-shot=true2023.10 | 0.632 | |
| UNITER w/ DCL*Zero-shot=true2023.10 | 0.62 | |
| UNITER(Large)Zero-shot=true2023.10 | 0.606 | |
| AudioCLIPZero-shot=true2023.10 | 0.6 | |
| CLIP w/ DCLZero-shot=true2023.10 | 0.584 | |
| Late Fusion w/ DCLZero-shot=true2023.10 | 0.577 | |
| CLIPZero-shot=true2023.10 | 0.563 | |
| Late FusionZero-shot=true2023.10 | 0.55 |