robotics
The Visual Plan Arrives in One Primitive
VFlash predicts complete points, boxes or trajectories in one forward pass instead of emitting their coordinates token by token.
Summary
VFlash predicts complete points, boxes or trajectories in one forward pass instead of emitting their coordinates token by token.
A lightweight drafter exploits the known structure of each visual primitive and adapts to its type, dimensionality and length. Across the evaluated 2D and 3D robot-trajectory tasks, the method reduces prediction error relative to autoregressive decoding. Preliminary single-rollout measurements show roughly 25-times faster 3D point-motion generation, while a conditioned world-action model reaches 70.2 percent average success across four MolmoSpaces suites versus 62.6 percent for the baseline.
Why it matters
VFlash predicts complete points, boxes or trajectories in one forward pass instead of emitting their coordinates token by token.
Limits and context
No additional limitation was separately recorded.
Key claims
VFlash predicts complete points, boxes or trajectories in one forward pass instead of emitting their coordinates token by token.
Evidence: source-2026-11-06-006
Sources
- arXiv preprint 2611.08449, original v1arXiv · primary research
Corrections
No corrections have been recorded for this story.