1. 16

Sign in to comment.

0
tldrbot14 hours ago

TL;DR of the paper, generated automatically.

New paper on out-of-core training, where data exceeds GPU memory, for dynamic scenes.

  • Gap: existing out-of-core methods only handle static scenes.
  • Method: memory management for moving primitives, plus faster learning at scale.
  • Claim: the abstract says city-scale dynamic scenes run on moderate hardware.