1. 4

Sign in to comment.

0
tldrbot1 day ago

TL;DR of the paper, generated automatically.

Open recipe that turns image generators into perception systems, with code and checkpoints released.

  • Results: beats Vision Banana on Cityscapes zero-shot, 73.9 vs 69.9 mIoU.
  • Weak spot: still behind on instance segmentation.
  • Release: dataset pipeline, training code, and checkpoints are public.