Photorealistic AI Generated Images

Just as a side note, what does "IV content" mean here?
IV as an Image Video. The solo outfit and posing fair like MMR-XX### or IMBD series from back a while. But as of now (at least on the platform I use) the clothes have to stay on, and cannot be as skimpy as I'd like. But I find ways to get some great content at least for the base images.

imbd-252-jpg.3355103


mmr-aa165-0-jpg.3585764
 
Last edited:
So here are 24 small 640x640 versions as an example. Depending on how responsive the AI is I can get some interesting actions. Turning around, bending over, squatting. I guess so long as the AI doesn't think it is too provocative or overtly sexual? Anyway, just to show what's possible and things out there will only get better and very quickly I bet. These are just 5 second clips, that's the default output for the basic credit cost per generation. I keep that default because it allows me a greater number of generations in case the AI gets things wrong, and it will. I sometimes go through 5 or 6 gens before I give up something I'm trying to do.

clips.jpg

AkibaShortShowcase.rar (36.3 Mb)
 
More improvements to my standalone model. Still has the 'A.I generated' look and feel but a vast improvement on things like hands and body proportions, and far better adherence to prompts.

Progressr_0133.jpeg
 

Attachments

  • Progress_0148.jpeg
    Progress_0148.jpeg
    122 KB · Views: 16
  • Like
  • Love
Reactions: jcjc and CoolKevin
I feel that the next huge step of AI would be including all the realistic details actually: dark circles under the eyes, stretch marks (you can have them in any age because of growth spurts), skin blemishes... We often turn to AI for something ideal, but then we realize that we actually want something real but in ideal conditions, we want to get as close as possible to recreating real life that we can control, not the ideal life... I guess... Or maybe I'm wrong. But just a thought :)
 
I feel that the next huge step of AI would be including all the realistic details actually: dark circles under the eyes, stretch marks (you can have them in any age because of growth spurts), skin blemishes... We often turn to AI for something ideal, but then we realize that we actually want something real but in ideal conditions, we want to get as close as possible to recreating real life that we can control, not the ideal life... I guess... Or maybe I'm wrong. But just a thought :)
I see that now on the gens of older ladies. I think you're right, it will only get better and more realistic.

arms.jpgbottom.jpg
 
anyone have '' image to text '' free online

i focus on actress s posing
 
Image to text? Or text to Image?
image to text ...

i want to upload the image on the net and then the a.i. give me " text '' so i can use this text (promt for a.i.) for generate new image
 
Hugging Face has a lot of different tools, you can start your search there (use the Search Spaces field; it's not very visible, but it's below the Spaces heading)
 
anyone have '' image to text '' free online

i focus on actress s posing
Not exactly 'online' but I use kohya_ss for creating the initial prompts for the models I've been training. It's a relatively quick and easy install, and can generate prompts from literally thousands of images in a couple of minutes.
 
Not exactly 'online' but I use kohya_ss for creating the initial prompts for the models I've been training. It's a relatively quick and easy install, and can generate prompts from literally thousands of images in a couple of minutes.
thank you so much
 
  • Like
Reactions: _2old_
any one try to do something like this ??

if you can do it ,,plz sharing me '' promt ''





473053818_1002613951892770_8916887898764048589_n.jpg473053818_1002613951892770_8916887898764048589_n.jpg


i dont like this at first ,,, its funny ..

some how i start to like it lolfe898953-1796-4f95-a2c0-60e6f4f81279.png
 
Last edited:
  • Like
Reactions: CoolKevin and jcjc
i am very bad on using forums function.

can i ask how long do you need for training these model ???

and how many people s face do you need for creating these ?? how many pic and how long

so i want to test it on line

because there is many option and tut on the net every day and you said you need 5 pics

can i ask what do you use for training now

flux,. lora or comui or the other way ????
I have trained all the models used for my posts in this thread. The majority are purely for private use so I've not shared them anywhere.
I am currently training a bunch of lora that are combinations of various Jav\Idols or European pornstars. I'm fine tuning these before I release them into the wild.

The lora below has the attractive working title of JavBod 3.5, a composite of several girls. It's almost ready ;)View attachment 3568721

(A few of my early experiments still lurk on Civitai somewhere, but lack a certain degree of polish. A bit of detective work will find them if you're interested).

I trained a very small Maria Ozawa model with only 5 pics. While the face likeliness isn't there, because I was using images from the same set, it picked up on her naked body rather brilliantly, including her magnificent pubes

View attachment 3570330

Maria gifted us this amazing body we could use with any face...
 
any one try to do something like this ??

if you can do it ,,plz sharing me '' promt ''





View attachment 3611435View attachment 3611435


i dont like this at first ,,, its funny ..

some how i start to like it lolView attachment 3611611

akiba1-2.jpg

I put the prompt in a spoiler:
1girl, Japanese, solo, 3d, bangs, barefoot, black hair, blue tshirt, striped boxer shorts, blurry, blurry background, blurry foreground, bookshelf, crossed arms, depth of field, indoors, lips, looking at viewer, lying, nose, on stomach, photo \(medium\), photorealistic, realistic, short hair, smile, solo, the pose, BREAK, a girl laying on the floor with her legs crossed and her hand on her chest, looking at the viewer, movie still, a photo, precisionism
Negative prompt: NSFW, Cleavage, Pubic Hair, Nudity, Naked, Au naturel, Watermark, Text, censored, deformed, bad anatomy, disfigured, poorly drawn face, mutated, extra limb, ugly, poorly drawn hands, missing limb, floating limbs, disconnected limbs, disconnected head, malformed hands, long neck, mutated hands and fingers, bad hands, missing fingers, cropped, worst quality, low quality, mutation, poorly drawn, huge calf, bad hands, fused hand, missing hand, disappearing arms, disappearing thigh, disappearing calf, disappearing legs, missing fingers, fused fingers, abnormal eye proportion, Abnormal hands, abnormal legs, abnormal feet, abnormal fingers
Steps: 30, Sampler: DPM++ 2M SDE, Schedule type: Karras, CFG scale: 6, Seed: 1194644121, Size: 1024x1024, Model hash: d61863bf3c, Model: cyberrealisticPony_v7, Clip skip: 2
 
1girl, solo, looking at viewer, smile, short hair, bangs, skirt, shirt, black hair, closed mouth, short sleeves, lying, barefoot, day, indoors, medium hair, mole, blurry, black eyes, feet, lips, looking to the side, toes, depth of field, blurry background, soles, leg up, bob cut, blue shirt, on stomach, t-shirt, realistic, nose, on floor, sliding doors, the pose
 

Attachments

  • image.png
    image.png
    1.2 MB · Views: 9