Yomino
Member
Historias de Yomino
-
Staff Applications are OPEN! [ Staff / Moderator ] More Info HERE Help us make a better forum for everyone!
[Yomino] Work
- Thread starter Yomino
- Start date
- body-modification feminization gender-change gender-morph transformation
yano2mch
Professional Geeky Perv
I've been playing with SD for a bit. I'm also new to it.
Probably the hardest part is getting consistency. Also you tend to have to get like 20 images to find one you're happy with, or do in-painting to fix problems.
On the pdf you released, i see differences mostly in his hoodie, the sleeves and string ends changing color or style in minor ways, slight differences in hair color, etc. Overall unless you're looking thankfully you won't notice.
If you have several images of a drawn individual or generated one you want, i think you can train a LORA and then get the consistency easier, but mostly for face/outfit, possibly locations too.
Another option would be using a model image and doing img2img to alter parts or the whole image keeping most of the outfits/postures the same while changing other things. Finding a sweet spot for what effects you want can take quite a bit of work and GPU cycles before you're happy.
Anyways curious to see what you might put up later.
parya
New member
I can tell you what kind of content I would buy.
I don't need pictures, but stories with pictures.
A young or adult man, for some reason, is mistaken for a woman or, better yet, a young girl. He is drawn into a gender intrigue, although he himself did not strive for this. Those around him gradually see him as a woman. And for some reason he cannot get out of this plot.
There must be a plot. If in the end the main character has everything he needs to get married and give birth and accept himself as happy - that's bingo!
And yes, I cannot be scared or offended by pictures with coercion or even aggression, self-harm. Pictures generally exist in a fictional world, and even in the real world, modern people usually have sex from the age of 14, so I don't like this nonsense about "don't draw children". To become a happy real woman, you need to live through all the stages from a girl to a teenager in one way or another.
In my opinion, stories should begin and end. For example, 1 story per week. There may be 10 pictures, maybe 40 - depending on how many are needed for the plot. The quality of the pictures and how many fingers the characters have do not interest me. The text can be written in any way you like. The story may have secondary plots or united characters - why not.
I don't need pictures, but stories with pictures.
A young or adult man, for some reason, is mistaken for a woman or, better yet, a young girl. He is drawn into a gender intrigue, although he himself did not strive for this. Those around him gradually see him as a woman. And for some reason he cannot get out of this plot.
There must be a plot. If in the end the main character has everything he needs to get married and give birth and accept himself as happy - that's bingo!
And yes, I cannot be scared or offended by pictures with coercion or even aggression, self-harm. Pictures generally exist in a fictional world, and even in the real world, modern people usually have sex from the age of 14, so I don't like this nonsense about "don't draw children". To become a happy real woman, you need to live through all the stages from a girl to a teenager in one way or another.
In my opinion, stories should begin and end. For example, 1 story per week. There may be 10 pictures, maybe 40 - depending on how many are needed for the plot. The quality of the pictures and how many fingers the characters have do not interest me. The text can be written in any way you like. The story may have secondary plots or united characters - why not.
parya
New member
And we shouldn't "sell a pig in a poke". Produce one or two full-fledged stories for free distribution, so that the buyer understands what he is signing up for. Maybe shortened without explicit scenes, but it should be clear that 4 dollars is not for several dozen pictures... ..first we will have to work on our reputation. The competition is big
parya
New member
I would reccomend you also do some research among the platforms and way how to overwalk its policy. As we can see here, Partion very strict to the publishing content.
parya
New member
Yes, many authors were banned last few years in "white" internet. Thats why you need to use patrion/fansly just to gather the money, but publish the real materials thru sending personal password for secured zip on dropbox/mega/google drive. Remain on patrion the "clean version" of it, without anything even close to be against the policy. From my POV the creative person can find the way how to express itself at 100% without being blocked.
yano2mch
Professional Geeky Perv
Personally, looks halfway interesting. However it feels like there's a lot of missing transition stages, are these all on the same day? Why are the clothes constantly changing? (Assuming it's on the same day) and the ] should be a ?
Lastly why did he turn into Tatsumaki? (Though seeing the filename tatsu transform, that kinda makes sense now so i guess that question can be dropped).
Otherwise, individually all the images look okay.
Yomino
Member
It doesn't really make much sense, I was just testing how the AI worked and researching features.Personalmente, parece medianamente interesante. Sin embargo, parece que faltan muchas etapas de transición. ¿Se dan todas el mismo día? ¿Por qué cambia la ropa constantemente? (Suponiendo que sea el mismo día) y el ] debería ser un ?
Por último, ¿por qué se convirtió en Tatsumaki? ( Aunque al ver el nombre de archivo tatsu transform , eso tiene sentido ahora, así que supongo que esa pregunta puede descartarse ).
Por lo demás, individualmente todas las imágenes se ven bien.
yano2mch
Professional Geeky Perv
Mhmm. Probably the hardest thing is getting consistency between frames, either of the same face/person or the clothing and area. At which likely img2img is more preferred, as you'd change an image to match what you want and it would retain a portion of the inferred data; layout, character positions, etc. Makes me wonder if the new age of AI comics would be using two or three people in outfits and taking pictures of old, then img2img and replace figure 1 with spiderman, figure 2 with wonder woman etc etc. Then crop them out and put a background premade or generated with their positions in mind.
Grimmlord1995
Active member
AI images are looking good but would be great if you could start off with some character info, so that instead of a random girl coming with needle, you could give some storyline with character details, you can make her into school bully or part of schools cult club looking for someone to run experiment on. Just a thought but keep up the good work.
