Physical Commonsense Reasoning On Physical
评估指标
Without Audio (Acc %)
评测结果
各个模型在此基准测试上的表现结果
模型名称 | Without Audio (Acc %) | Paper Title | Repository |
|---|---|---|---|
| UNITER (Large) | 60.6 ± 2.2 | PACS: A Dataset for Physical Audiovisual CommonSense Reasoning | |
| Human | 90.5 ± 3.1 | PACS: A Dataset for Physical Audiovisual CommonSense Reasoning | |
| Merlot Reserve (Large) | 68.4 ± 0.7 | PACS: A Dataset for Physical Audiovisual CommonSense Reasoning | |
| Majority | 50.4 | PACS: A Dataset for Physical Audiovisual CommonSense Reasoning | |
| CLIP/AudioCLIP | 56.3 ± 0.7 | PACS: A Dataset for Physical Audiovisual CommonSense Reasoning | |
| Late Fusion | 52.5 ± 1.6 | PACS: A Dataset for Physical Audiovisual CommonSense Reasoning |
0 of 6 row(s) selected.