KREA has a "
Realtime" fast generation feature on their website -- and the KREA trained AI models have to be specific for Realtime (
which will not work in their FLUX generation feature). So I trained a "Realtime" AI model on 14 long pose drawings, drawn mostly from the
end of my book --
KRRRL Long Drawings:
Uploading 14 long pose images
The "Realtime" model
was the last of the 5 free AI models
that KREA trained on my drawings
REALTIME
Sitting Variations
What is the difference between my original drawings, and those drawings altered by
KREA AI? Below are two comparisons:
I uploaded two high resolution images of my long post drawings of sitting women, to the Realtime feature of KREA, and using the Realtime model trained on my drawings -- KRRRL Long Drawings:
SITTING LADY 1
to make AI variations in KREA Realtime
SITTING LADY 2
I uploaded a different
high resolution image of my drawing of a sitting woman, to the Realtime feature of
KREA, and using the Realtime model trained on my drawings --
KRRRL Long Drawings. I generated new variations -- including a few clothed figures:
to generate new AI variations in KREA Realtime
OTHER KREA AI IMAGES
based on my figure drawings
My drawing in the style of Bernini
Midjourney did not censor the "clothed" version above, so I Blended that image with another
Rope figure that I generated in KREA:
The previous image blended in Midjourney
No comments:
Post a Comment
Note: only a member of this blog may post a comment.