AI Tools & Models HOTFeaturedAll
arXiv cs.AIScore60

Training-Free Acceleration of Multi-View 3D Transformers via Regime-Aware Attention Approximation

Summary

This work accelerates multi-view 3D transformers without training, speeding up scene reconstruction from unposed images.

Why this matters

Relevant to this topic

Open original ↗

Source: arXiv cs.AI · arxiv.org

Related items