Here is a tutorial on how to operate VQGAN+CLIP by Katherine Crowson! No coding knowledge necessary.
I can't guarentee I'll see all the requests we get, so if you want a surefire way to have your image generated, make it yourself! It's super fun.
Conversation
Hi!!!!! Thank you so much for this tutorial! I'm having a blast! I have a quick question. I noticed you said you can put multiple image file names in the "target image" section. How do you separate the file names to do that? Is it with the same "|" character? :)
1
1
6
Show replies
Hi! How many iterations do you recomend for good results? or how many iterations are you using? I am using it but I feel that my images are not weird enough
1
4
I'd say anything past 2000 is probably overkill unless you want it to be *really* detailed. Usually it will have taken some kind of recognizable form by 500.
2
15
Show replies
The largest dimensions I could feed it without getting an out of memory error are 300x300... is there something I'm doing wrong, or is it just temperamental?
1
3
That's up to the arcane ways of Google Collab. I think it dedicates memory based on what GPUs it has available and how much memory you've used historically. I suspect the sudden influx of people trying it out from this post might have clogged it some :/
1
7
Show replies
Hi, thank you so much for this tutorial! I'm running into an error at the "Carga de bibliotecas y definiciones" cell - it says there's no module found named 'taming,' and fails to run. Any fixes?
1
1
Show replies
Great tutorial! Just wanted to point out this area in the introduction needs a correction:
and are two different (and wonderful!) people
1
2
12
Oh my goodness! Thanks for letting me know! I think Glen had a link to Advadnoun's Patreon in their Youtube descriptions, which led me to assume they were one and the same. This is excellent, I'll correct it now
11