Preprint
Oct 2026
BlazeVGGT: 13x Faster VGGT at 2,000 Frames via Camera Token without Training
BlazeVGGT is introduced, a training-free framework for structure-aware frame-level key--value sparsification that preserves dense per-frame predictions while eliminating redundant cross-view computation, allowing the method to be directly integrated into pretrained VGGT as a plug-and-play module without parameter updat...
Shu-Jun Yao, Jun-Hui Lin, Cong-Ying Han et al.
· 1 citation