VLA-Δ: Learning Physical State Transition for Vision-Language-Action Models
Open paper
Benchmark
Model
Rank
Results
robot-manipulation-on-robotwin-2-0-easy-50-tasks
VLA-Δ
#20
Success Rate: 90.06
robot-manipulation-on-robotwin-2-0-hard-50-tasks
VLA-Δ
#20
Success Rate: 89.74
robot-manipulation-on-libero
VLA-Δ
#40
Success Rate: 98.4
Rank counts only results with a code link.