Grasping with orientation variation on Take Kettle
96Grasp Success RateMSARNN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MSARNNInput Image=Stereo, Visual Encoder Backbone=MSA2026.05 | 96 | 92 | |
| MSARNNInput Image=Stereo, Visual Encoder Backbone=SA2026.05 | 76 | 70 | |
| DPInput Image=Stereo, Visual Encoder Backbone=ResNet2026.05 | 52 | 4 | |
| ACTInput Image=Stereo, Visual Encoder Backbone=ResNet2026.05 | 22 | 6 | |
| π0Input Image=Stereo, Visual Encoder Backbone=ViT, Parameters=3.5B2026.05 | 0 | 0 | |
| SmolVLAInput Image=Stereo, Visual Encoder Backbone=ViT, Parameters=0.45B2026.05 | 0 | 0 | |
| MSARNNInput Image=Mono, Visual Encoder Backbone=MSA2026.05 | 0 | 0 |