After a long break, i'm trying to return to creating galleries, if producing AI slops can even be called creation.
The reason i gave up on implementing my scripts was that i couldn't generate the scene the way i wanted, or couldn't create anything that even remotely resembled the description that was entered in the prompt window. However, i recently started learning controlnet, and the first results give me some optimism. It seems that the problem with the stubborn AI model now has been solved, and scenes that are particularly difficult to generate can somehow be generated. At the same time, it is worth to note that it is still difficult to do, because the controlnet i use accepts sketches made with a screenless tablet (wakom or something like that) as the input. I do have the tablet, but that's just half of the whole thing: i also need to learn how to draw at least sketches.
Here's a little photoshoot of Nakamichi Mayuri that's *not* part of the main narrative. It's like a special chapter in a manga, but it's not a manga (and it's not a chapter). The main reason for creating this gallery is to inform others that my creative stagnation is gradually being overcome.
The last two pictures are my godlike sketch and how it turned out after being interpreted by the controlnet and WAI-NSFW-illustrious_v.14.