search

lucidrains / deep-daze

Simple command line tool for text to image generation using OpenAI's CLIP and Siren (Implicit neural representation network). Technique was originally created by https://twitter.com/advadnoun
MIT License
4.37k stars 327 forks source link

Doesn't save any images to disk #62

Open n00mkrad opened 3 years ago

n00mkrad commented 3 years ago

The current version (pip) doesn't save any images, no matter which save_every value.

Tested locally (Win 10) and on the simplified Colab notebook.

lucidrains commented 3 years ago

@n00mkrad ohh, for some reason it is hidden behind a CLI flag now --save-progress, but i defaulted it to true in the latest version! https://github.com/lucidrains/deep-daze/releases/tag/0.6.3

afiaka87 commented 3 years ago

Nice I think that was my bad. I always use save-progress.

n00mkrad commented 3 years ago

I mean the CLI can't display images, I don't see the point of it being disabled by default lol

lucidrains commented 3 years ago

@n00mkrad I think @afiaka87 wanted to only save it at the very end

afiaka87 commented 3 years ago

Yeah I believe that's how it turned out. Ultimately just not testing with the CLI (I'm using the notebook) as I dont have a great GPU locally and not testing all the parameters before shipping. My bad.