Self-collected
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Place Recognition | Self-collected dataset | AR@180.78 | 11 | |
| 3D Surface Reconstruction | self-collected datasets | Accuracy14.65 | 10 | |
| Value-Order Correlation | Self-collected 3 embodiments, expert demonstrations | Value-Order Correlation (VOC)0.99 | 7 | |
| Video Generation | Self-collected (test) | Human Consistency3.52 | 6 | |
| Video Animation | Self-collected (test) | FID (Fréchet Inception Distance)51.6 | 6 | |
| Localization | Self-collected Seq. 3 | Horizontal Position RMSE (m)0.94 | 5 | |
| Localization | Self-collected Seq. 2 | Horizontal Position RMSE (m)0.49 | 5 | |
| Localization | Self-collected Seq. 1 | Horizontal Position RMSE (m)0.5 | 5 | |
| UAV Tracking | Self-Collected | Precision89.7 | 5 | |
| Long-horizon streaming video generation | self-collected e2e dataset (test) | FID12.01 | 4 | |
| Depth Estimation | Self-collected | Abs Rel0.185 | 3 | |
| Sentiment Classification | Self-Collected | Accuracy76.01 | 1 | |
| Parkinson's Disease Classification | Self-Collected | Accuracy92.6 | 1 |
Showing 13 of 13 rows