Table 1 Comparison with transformer. The best results are boldfaced.
From: Hierarchical Information-guided robotic grasp detection
Approaches | Accuracy (100%) | Complexity | ||
|---|---|---|---|---|
Image-wise | Object-wise | Params (M) | FLOPs (G) | |
Transformer | 97.88 | 96.52 | 68.25 | 73.91 |
Ours | 98.91 | 97.30 | 1.57 | 15.14 |
