Illustrious version NOW AVAILABLE.
As someone who does real art and who has also fiddled around with AI. I'd like to say something.
The effort it took to train this model is nothing compared to the effort it takes to learn, study, and master art. Anything created with this model should be created only for novelty and not for profit or validation. As it takes about 10-60 seconds to generate an image with this, Anything created with this model is inherently worthless. No amount of epochs, samples, batch sizes etc will even replace genuine art.
You aren't special or talented for using AI. The results might look good, but it's just for fun. Please support real artists.
Trained on 500 well tagged images. The images were of the pose from many different angles, art styles, mediums, etc... The tagging method involved tagging everything except what is essential to the pose. Leaving out tags referencing the pose itself, unless I wanted it to be a variable.
I hope this will be a huge improvement over the last model. It is very flexible in terms of characters, clothing, poses and camera angles. This model produces images of two characters doing a reverse headscissor.
All the preview images are created without inpainting or control net.
After a bit of testing, the model works well on weights 0.4-1. the concept is pretty well understood from just the trigger word, but prompting can help it show up if it's not giving you good results.
Trigger word: revscs
Prompt Guide (variables within the model):
from side, from front, from behind, close up, cowboy shot, feet out of frame, full body, holding another's wrists, standing, on side, on ground, twisted torso, looking back, groping, ass grab, grabbing another's ass
Helper Prompts (not necessary but will help push the model in the right direction):
69, head between thighs, sitting on face, face in ass, lying on person, lying on ground
Negative Prompt Guide:
fat, muscular, thick, wide hips, fat ass, thick thighs
(the model has a bias towards outrageous female lower body proportions so you may have to prompt it towards something normal)
Please post your image generations with this model. They help training future models.
If you found this model useful, please support me on fanbox: https://arturitwo.fanbox.cc/
Description
Just retrained the same dataset on Illustrious, not sure if it works better or not
FAQ
Comments (2)
As much as I agree with the sentiment that things created with AI art are "worthless", I'm sure classical painters thought the same when Photoshop came out and people could undo mistakes, work in layers and chop n' change at will, without creating anything physical in the process.
No matter what we do, technology will always progress and make things quicker and easier, to the point where we wake up, press a button and the human experience is complete in one fancy package.
Tempted to tell bro to put the fries in the bag don't nobody gaf