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
4.38k stars 327 forks source link

CLI is broken - "missing 1 required positional argument: 'name'" #51

Closed afiaka87 closed 3 years ago

afiaka87 commented 3 years ago

Usage:

imagine "phrase"

Output:

imagine "tacos"
TypeError: __init__() takes 1 positional argument but 2 positional arguments (and 18 keyword-only arguments) were given
afiaka87 commented 3 years ago

I've fixed the bug (and added the ability to choose between the ViT-B/32 or RN50 models of CLIP) in this pull request.

https://github.com/lucidrains/deep-daze/pull/52

afiaka87 commented 3 years ago

Having a little trouble with the CLIP model code. Here's a PR with just the CLI fix (it's one line in cli.py).

https://github.com/lucidrains/deep-daze/pull/53

NotNANtoN commented 3 years ago

Hey, I'm sorry. I introduced this because I changed the text to be optional in my PR. I did not test the CLI and thought someone would check it. It surprised me that the PR was merged directly - but now I know. Next time I'll check more thoroughly!

afiaka87 commented 3 years ago

Hey, I'm sorry. I introduced this because I changed the text to be optional in my PR. I did not test the CLI and thought someone would check it. It surprised me that the PR was merged directly - but now I know. Next time I'll check more thoroughly!

No problem. The versioning probably mitigated it to a degree. Having said that, @lucidrains it might be good to remove the pip --upgrade call in your colab notebook.