Vision-based single-gate
Benchmarks
TaskDatasetSOTA resultTrendResults
Vision-based single-gate Compositional Generalization v1 (Held-out)
59.4Success Rate
2
Vision-based single-gate Compositional Generalization v1 (train)
69.8Success Rate
2
Vision-based single-gate Venue Transfer - Pool 0-shot v1 (test)
36.7Success Rate
2
Vision-based single-gate Venue Transfer 3 Venues v1 (train)
63.3Success Rate
2