RAP
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Data Extraction Attack | RAP | Attack Success Rate (ASR)100 | 32 | |
| Pedestrian Attribute Recognition | RAP v2 | mA87.4 | 27 | |
| Pedestrian Attribute Recognition | RAP | mA81.87 | 26 | |
| Attribute Recognition | RAP zero-shot | mA81.09 | 18 | |
| Pedestrian Attribute Recognition | RAP v1 | Accuracy72.23 | 15 | |
| Data Extraction Attack on Agent Memory | RAP (test) | Equality Score71 | 12 | |
| Pedestrian Attribute Recognition | RAP existing vs. zero-shot v2 (multiple) | mA78.21 | 8 | |
| Pedestrian Attribute Recognition | RAP-OV (test) | F1@348.9 | 6 | |
| Person Re-identification | RAP Language 1.0 (test) | DPD5.3 | 5 | |
| Person Re-identification | RAP Ethnicity 1.0 (test) | DPD7.4 | 5 | |
| Person Re-identification | RAP Gender 1.0 (test) | DPD0.4 | 5 | |
| Person Re-identification | RAP Race 1.0 (test) | DPD2.2 | 5 | |
| Recognition | RAP Multi Concept 1 Reference View 11 | Precision100 | 3 | |
| Pedestrian attribute recognition | RAP v2 (test) | mA82.34 | 3 |
Showing 14 of 14 rows