Robot Manipulation On Rlbench
평가 지표
Input Image Size
Succ. Rate (18 tasks, 100 demo/task)
평가 결과
이 벤치마크에서 각 모델의 성능 결과
비교 표
| 모델 이름 | Input Image Size | Succ. Rate (18 tasks, 100 demo/task) |
|---|---|---|
| perceiver-actor-a-multi-task-transformer-for | 128 | 1.3 |
| autoregressive-action-sequence-learning-for | - | 86.0 |
| coarse-to-fine-q-attention-efficient-learning | 128 | 20.1 |
| act3d-infinite-resolution-action-detection | 256 | 65 |
| auto-lambda-disentangling-dynamic-task | - | - |
| rvt-robotic-view-transformer-for-3d-object | 128 | 62.9 |
| towards-generalizable-vision-language-robotic | 256 | 83.1 |
| perceiver-actor-a-multi-task-transformer-for | 128 | 49.4 |
| perceiver-actor-a-multi-task-transformer-for | 128 | 42.7 |
| instruction-driven-history-aware-policies-for | - | 45.3 |
| polarnet-3d-point-clouds-for-language-guided | 128 | 46.4 |
| sam2act-integrating-visual-foundation-model-1 | 128 | 86.8 |
| 3d-diffuser-actor-policy-diffusion-with-3d | 256 | 81.3 |
| perceiver-actor-a-multi-task-transformer-for | 128 | 1.3 |
| rvt-2-learning-precise-manipulation-from-few | 128 | 81.4 |