VLA-3D
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Robotic Spatial Reasoning | VLA-3D Target never directly observed (n = 10) | Mean Information Gain (nats)1.65 | 5 | |
| Robotic Spatial Reasoning | VLA-3D Target observed by vision | Mean IG (nats)3.77 | 5 | |
| target object grounding | VLA-3D | Accuracy99.58 | 5 |
Showing 3 of 3 rows