The Better You Learn, the Smarter You Prune: Towards Efficient Vision-Language-Action Models Via Differentiable Token Pruning
Open paper
Benchmark
Model
Rank
Results
robot-manipulation-on-libero
LightVLA (LIBERO average)
–
Success Rate: 97.4
robot-manipulation-on-libero-goal
LightVLA (LIBERO-Goal)
–
Success Rate: 98.2
Rank counts only results with a code link.