The goal is to generate intriguing IMAGES with AI
"..so the thing becomes an exercise in trying to force the system to do something that is actually against it's nature...because it's nature is to suck up and take the most probable next word...no...I WANT THE LEAST PROBABLE NEXT WORD..."
LOW EFFORT NOTHINGS
...but apparently I only generate "Low Effort Nothings" with my AI ART:
SUMMARY POSTS
about AI IMAGE ART 2024
.png)
AI ART
IMAGE SOFTWARE
Added a Universal Upscaler in March
Canva acquired Leonardo.AI at the end of July
A program written by WEBSIM on July 2, 2024
HUGGING FACE:
AI ART
GOSSIP
The AI website I liked the most is Leonardo.AI -- because they let me train a LoRA with my drawings, so that I can generate decent FAKE drawings in my KRRRL style (the "KrrrlLong" model works the best). However once they went to a paid program, this does not seem to work as well as it used to.
Leonardo.AI let us use their new "Alchemy" option for free for one week (January 22 - 28, 2024), and it worked marvelously, especially for blending my style with that of the sculptor Bernini.
Then at the end of July 2024, Canva bought Leonardo. AI -- and now it censors, suggesting that I need to use the paid version to upload my figure drawings.
MIDJOURNEY is recognized as the most powerful online AI ART program out there -- however they are overzealous in censoring. They reject most of my figure drawings when I try to upload them. And after I made a lot of "Moodboards" at the end of December 2024, they censored everything I made with a "Moodboard," even benign cat images.
Midjourney did not release a new version in 2024. However they did release a lot of new features which I covered in this blog post: AI IMAGES 2024 -- Evolution of Midjourney
DEEP DREAM GENERATOR keeps releasing new features -- the one I like the most is the "AI UPSCALER" which they released about June 4, 2024. It does not censor my figure drawings. It does tend to shoehorn the results into one type of figure -- but it is the "AI UPSCALER," and is not suppose to go too much off into the deep end:
RECRAFT surfaced originally as an AI ART program that generated vectors back in May. Then since October 30th they surfaced as a superior AI ART generator:
.png)
I like the way rope can curl around a shape, not unlike the way my lines describe a human figure. Moreover manipulating rope does not offend the censors, nor is it NSFW. Therefore I like to push the rope theme through AI, and I am pleased with the results:
.png)
GOOGLE
I asked Google to "generate a figure drawing in the style of Krrrl" and it made four generations -- you have to turn on "SGE" in Google to make it work:
Then in December I just chatted with GEMINI AI to generate images:
Generated by chatting in GOOGLE GEMINI
IDM-VTON puts outfits on images one uploads, on Hugging Face:
Sakana AI reinterprets images as Japanese style, on Hugging Face:
Generated with Sakana AI
FACE POKE can shift the gaze of a portrait, on Hugging Face:
Generated in Face Poke
Image Mixer does not blend images as well as Midjourney...and no longer works on Hugging Face:
.jpg)
No comments:
Post a Comment
Note: only a member of this blog may post a comment.