I have to get a handle on that program - just noodled around so far. Most stuff I see on VQGAN+CLIP is pretty poor. You seem to have mastered how to use that AI.
The look of it echoes Beksinski.
You are viewing a single comment's thread from:
I have to get a handle on that program - just noodled around so far. Most stuff I see on VQGAN+CLIP is pretty poor. You seem to have mastered how to use that AI.
The look of it echoes Beksinski.
Mastered would be saying too much... I think is the AI who mastered how to fool me hehe. Jokes apart, everything is about to feed the machine properly. Poor results comes from non sensical requests. if I ask for a very detailed painting of a Ferrari car, by Picasso.. as he never painted such thing, and his style is not detailed, not even realist... 99% of times, the output of the AI gonna be poor. At end it is about use the common sense, and understand how the AI works.
You can start with the original VQGAN Notebook, although sadly, it is in Spanish language. There are more complex notebooks around, feel free to ask me for them and resources about this topic.
Here the original notebook: VQGAN+CLIP
And yeah, you got me haha. Zdzislaw Beksinski, is one of the main ingredients that I have I used for this recipe. It seems that adding Caravaggio to him, gave to the result more "volume"... I don't know if it is the proper word to define what I mean, at end I'm not an artist.
Thanks a lot for your comment, and grab this slice of !PIZZA too!
$PIZZA@thermoplastic! I sent you a slice of on behalf of @jotakrevs.
Did you know you can earn $VFT through the PIZZA farm? (1/10)