heystack

Introduction to VQGAN+CLIP

Here is a tutorial on how to operate VQGAN+CLIP by Katherine Crowson! No coding knowledge necessary.

From the heystacks.com archiveUpdated 85K views
Open original

Published history

Every published version keeps its original contributor and change message.

  1. Current version

    Imported legacy reference

    Legacy contributor · 2026-10-12

    Read exact source · Restore by submitting this source as a new revision.

    Read this version
    Open original Google document

Discussion 20

0/5,000

From the heystacks.com archive

  1. Ed

    This is such a user friendly guide, I'm so impressed! Though I've run into a bit of an issue - it can't seem to connect to the site to download Katherine Crowson's diffusion model - I'm guessing this means that the other models I downloaded are pretty much useless. Do you know if there's another mirror, or whether it's been taken down? Thanks!

  2. J

    This is absolutely amazing. I am getting mind blowing results. Do I own my creations?

  3. cdammr

    I can't for the life of me get this to run! First "NameError: name 'texts' is not defined" and then when I manually enter in the texts variable, I get "NameError: name 'args' is not defined" and I have no idea what else to do. I must be doing something wrong?

  4. Gabriele

    I keep having this issue in several cells: NameError: name 'abs_root_path' is not defined. Any tip?

  5. ME

    I can't seem to get coco, s-flckr, or faceshq to run every time I try it. Have you gotten those to work?

  6. Leo

    Is it legal to sell this art or is it copyright?

  7. J

    Copy the path

  8. Octipus

    Hi! Is there any way i can use an image as a starting point?

  9. Jwalk

    I don't understand how the Advanced VQGAN+CLIP Parameters impact the output. l I have played with them a but, but cant seem to figure out what they do. looking for documentation, but no luck yet.

  10. Jwalk

    I also was not able to download the diffusion model. Messaged the notebook author on twitter and he said he was looking for a mirror, but that was a few weeks ago, so not sure the status. Have been able to run other models though, just cant do clip guided diffusion.