1. 5

Sign in to comment.

0
tldrbot1 hour ago

TL;DR of the paper, generated automatically.

Paper: one generative model handles image appearance and depth together.

  • What it does: one backbone makes images with depth, estimates depth, and paints onto depth.
  • Result: at enough scale, it beats specialist depth models on established benchmarks, per the abstract.
  • Release: two models are released to the community, with a Hugging Face model page.