Deformable-object interaction on Retrieve Clothing
84Grasp Success RateMSARNN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| MSARNNInput Image=Stereo, Visual Encoder Backbone=MSA2026.05 | 84 | 78 | |
| π0Input Image=Stereo, Visual Encoder Backbone=ViT, Parameters=3.5B2026.05 | 76 | 76 | |
| DPInput Image=Stereo, Visual Encoder Backbone=ResNet2026.05 | 60 | 54 | |
| ACTInput Image=Stereo, Visual Encoder Backbone=ResNet2026.05 | 58 | 56 | |
| MSARNNInput Image=Mono, Visual Encoder Backbone=MSA2026.05 | 48 | 44 | |
| SmolVLAInput Image=Stereo, Visual Encoder Backbone=ViT, Parameters=0.45B2026.05 | 32 | 32 | |
| MSARNNInput Image=Stereo, Visual Encoder Backbone=SA2026.05 | 28 | 28 |