You are viewing a single thread.
View all comments
5 points

Looks interesting, but also somewhat complicated. Now generation has to run through 3 models? I thought it was bad enough having a separate refiner model for sdxl.

permalink
report
reply
4 points

It’s based on Würstchen which has real savings on training, but the VRAM requirement is through the roof.

permalink
report
parent
reply
2 points

Sdxl was 3 models too. Base, refiner and vae.

Cascade is 3. Lores, hires, vae.

permalink
report
parent
reply

Stable Diffusion

!stable_diffusion@lemmy.dbzer0.com

Create post

Discuss matters related to our favourite AI Art generation technology

Also see

Other communities

Community stats

  • 149

    Monthly active users

  • 851

    Posts

  • 1.7K

    Comments

Community moderators