VFlash: Fast, Accurate Visual Planning via Flash Primitive Decoding
VFlash: Fast, Accurate Visual Planning via Flash Primitive Decoding
This research introduces VFlash, a method for fast and accurate visual planning in physical AI tasks like robotics and autonomous driving. By exploiting the structure of visual primitives, VFlash predicts coordinates jointly using a lightweight drafter, supporting compact boxes and variable-length trajectories without requiring exact token sequence reproduction. Evaluated on 2D and 3D robot trajectory tasks, VFlash reduces prediction errors relative to autoregressive decoding. Preliminary measurements indicate approximately 25x faster generation for 3D point motion forecasting in single-rollout scenarios. A world action model conditioned on VFlash plans achieves 70.2% average success across four MolmoSpaces task suites, compared with 62.6% for the baseline.
The paper introduces VFlash, a method using Flash Primitive Decoding to jointly predict visual plan coordinates for robotics and autonomous driving tasks, reporting approximately 25x faster generation in single-rollout measurements and improved success rates on MolmoSpaces task suites compared to autoregressive baselines.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
Reported numbers
generation speedup
25
View original evidence
pproximately 25x faster generation for 3D point motion forecasting. A world action model conditioned on VFlash plans achieves 70.2% average success across four MolmoSpaces task suites, compared with 62.6% for the baseline.
Open source S5forward pass count
1
View original evidence
We introduce VFlash, a method for fast, accurate visual planning through Flash Primitive Decoding. VFlash exploits the known structure of visual primitives to predict their coordinates jointly, using a lightweight drafter that generates an entire primitive in one forward pass. The layout adapts to the primitive's type, dimensionality, and length, supporting compact boxes and long, variable-length trajectories. Cruci
Open source S3
- paper: not_reported
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
pproximately 25x faster generation for 3D point motion forecasting. A world action model conditioned on VFlash plans achieves 70.2% average success across four MolmoSpaces task suites, compared with 62.6% for the baseline.
Open source S5
ally, spatial accuracy need not require reproducing an autoregressive model's exact token sequence. VFlash therefore trains on annotated primitives and directly accepts structurally valid predictions without target-token agreement. Across the evaluated 2D and 3D robot trajectory tasks, VFlash consistently reduces prediction errors relative to autoregressive decoding. Preliminary single-rollout measurements indicate a
Open source S4
We introduce VFlash, a method for fast, accurate visual planning through Flash Primitive Decoding. VFlash exploits the known structure of visual primitives to predict their coordinates jointly, using a lightweight drafter that generates an entire primitive in one forward pass. The layout adapts to the primitive's type, dimensionality, and length, supporting compact boxes and long, variable-length trajectories. Cruci
Open source S3
Source：arXiv Robotics — research abstracts · arxiv.org