Imitation Learning on LIBERO
0.076Spatial MSELAOF-Action
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| LAOF-ActionAction supervision=true, Action ratio=1%2025.11 | 0.076 | 88.2 | 0.064 | 95.9 | 0.081 | 88.6 | 0.068 | 63.7 | 0.096 | 11.5 | |
| LAOM-ActionAction supervision=true, Action ratio=1%2025.11 | 0.108 | 86 | 0.09 | 91.1 | 0.127 | 86.3 | 0.086 | 61.6 | -0.066 | 8.7 | |
| LAOF2025.11 | 0.111 | 82.5 | 0.082 | 85.3 | 0.118 | 87.2 | 0.088 | 52 | 0.069 | 4.2 | |
| LAPO2025.11 | 0.162 | 80.4 | 0.139 | 81.2 | 0.219 | 84 | 0.154 | 44.7 | 0 | 0 | |
| CoMo w/ OFOptical flow constraints=true2025.11 | 0.172 | 76.2 | 0.129 | 89.7 | 0.113 | 82.6 | 0.139 | 57.9 | -0.03 | 4 | |
| CoMo2025.11 | 0.181 | 74.1 | 0.125 | 87.6 | 0.221 | 80.8 | 0.153 | 49.9 | 0.002 | 0.5 |