| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Activity Recognition | RealWorld | Accuracy55.5 | 18 | |
| Human Activity Recognition | RealWorld | F192.11 | 14 | |
| Visual Question Answering | RealWorld | Accuracy68.89 | 10 | |
| Domain-Incremental Human Activity Recognition | RealWorld | FA76.05 | 10 | |
| Visual instruction tuning | RealWorld | Score74 | 6 | |
| Blind Super-Resolution | RealWorld60 | ManIQA0.6333 | 6 | |
| Human Activity Recognition | RealWorld (test) | Weighted F194.88 | 5 | |
| Navigation | Realworld Navigation | Episode Accuracy68 | 4 | |
| Human Activity Recognition | RealWorld Full-Shot | Accuracy84 | 4 | |
| Manipulation | Realworld Manipulation Pick and Place | Episode Accuracy76 | 2 | |
| Activity Recognition | Realworld 43 (external evaluation) | Accuracy81.3 | 2 |