Skip to content

Comment on Generating transparent images using Stable Diffusion XL

Comments

reactions

1 - the way the dog at the end gets a reflection off the floor is pretty nice.

2 - i wonder how this compares in terms of latency/complexity with a comfyui pipeline that just does a typical edge detection/masking layer to achieve the transparency effect. however i dont think that method would work with the glass example as shown

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.