AI

kazuma · 06-Dec-23, 05:02 PM

I am trying to make AI wrestling content, but I am really new to the field. If anybody has any recommendations, I would love to learn more. Here is some content I made, I mostly based the girl on Skylar Rene and Jolene Hexx, but I am really new to the subject, so please forgive my shortcomings.

link to pics:
https://gofile.io/d/2EKNvf

edfan20 · 07-Dec-23, 08:42 PM

i loaded the new stable diffusion and asked it some ... specific things.. it's absolutely wild how far this has come.. i can't believe my computer made this out of thin air.

edfan20 · 07-Dec-23, 08:45 PM

one other that came out nicely

kazuma · 09-Dec-23, 08:15 AM

Did you see the video creation tools they just released? That is wild.

Goldeneyed · 09-Dec-23, 03:47 PM

Which software are you using?

kazuma · 09-Dec-23, 05:00 PM

Quote from: Goldeneyed on 09-Dec-23, 03:47 PMWhich software are you using?

I use:
https://github.com/AUTOMATIC1111/stable-diffusion-webui

tutorial:
https://www.youtube.com/watch?v=onmqbI5XPH8

edfan20 · 11-Dec-23, 04:51 PM

That's what i'm using as well.. there are some other tools that may be better though..

As of now- i'm trying stuff out with some of my favorite scenes, maybe swapping characters around.. it's not really possible to feed a prompt like "girl puts a wimpy guy in a headscissors" haha.. but I have a sense, for better or worse, that's right around the corner..

any way, added one more pic here, which I'm guessing some might recognize from the original- hope you don't mind Kazuma. it's been a fun thing to mess with.

kazuma · 13-Dec-23, 08:37 AM

Quote from: edfan20 on 11-Dec-23, 04:51 PMThat's what i'm using as well.. there are some other tools that may be better though..

As of now- i'm trying stuff out with some of my favorite scenes, maybe swapping characters around.. it's not really possible to feed a prompt like "girl puts a wimpy guy in a headscissors" haha.. but I have a sense, for better or worse, that's right around the corner..

any way, added one more pic here, which I'm guessing some might recognize from the original- hope you don't mind Kazuma. it's been a fun thing to mess with.

Your contributions are most appreciated, sir.

Tigigr · 29-Jan-24, 07:45 AM

Attempted to do some IA images, and even when managing to do something, it's pretty far to actually cool wrestling scenes and images... but, better than nothing, I think. Sorry for them not being at least very good, but just thought it would be interesting to share it here.

I had better results with punches and holds into the men's arms and shoulders, but not too much with chokes and generated girls image flexing her arms in the sports bra and etc to serve as 'this is the girl who defeated the guy' effect. But despite all the limitations, it's fun sometimes to try to generate them, specially now when I'm facing a leukemia

I still have a couple of images to post (including some other holds). I'll try to share in next ones.

dark87dan · 29-Jan-24, 10:32 AM

Quote from: Tigigr on 29-Jan-24, 07:45 AMAttempted to do some IA images, and even when managing to do something, it's pretty far to actually cool wrestling scenes and images... but, better than nothing, I think. Sorry for them not being at least very good, but just thought it would be interesting to share it here.

I had better results with punches and holds into the men's arms and shoulders, but not too much with chokes and generated girls image flexing her arms in the sports bra and etc to serve as 'this is the girl who defeated the guy' effect. But despite all the limitations, it's fun sometimes to try to generate them, specially now when I'm facing a leukemia

I still have a couple of images to post (including some other holds). I'll try to share in next ones.

Thanks. I always appreciate art and I think I'm going to try AI to create some of my favorite holds or even enhance a story.

Tigigr · 30-Jan-24, 06:12 AM

Quote from: dark87dan on 29-Jan-24, 10:32 AM
Quote from: Tigigr on 29-Jan-24, 07:45 AMAttempted to do some IA images, and even when managing to do something, it's pretty far to actually cool wrestling scenes and images... but, better than nothing, I think. Sorry for them not being at least very good, but just thought it would be interesting to share it here.

I had better results with punches and holds into the men's arms and shoulders, but not too much with chokes and generated girls image flexing her arms in the sports bra and etc to serve as 'this is the girl who defeated the guy' effect. But despite all the limitations, it's fun sometimes to try to generate them, specially now when I'm facing a leukemia

I still have a couple of images to post (including some other holds). I'll try to share in next ones.

Thanks. I always appreciate art and I think I'm going to try AI to create some of my favorite holds or even enhance a story.

Indeed. Would be awesome if the AI didn't had many limitations on doing arm chokes

, but I guess there's a chance for you to achieve your fav words.
Here's more images, i almost managed to create a rear naked choke and a guillotine (i think i still have more to post)

jynn · 30-Jan-24, 09:57 AM

IF you want interaction between people, are if you want some type of control over poses..you need to look into controlnet or t2i adapters. (t2i adapters are better, faster) This will allow you to apply different models like canny, depth, etc...to prompt the generative AI on how to use the noise to generate the image. Lastly, if you want to really get serious, you can train your on model on Kohya, teaching the AI what a headscissor looks like. (Time consuming, and you need good hardware).

As for other AI generators, here are some that I use... I no longer use A1111 as it has be more of a pain these days:
Fooocus: https://github.com/lllyasviel/Fooocus
ComfyUI (my daily driver): https://github.com/comfyanonymous/ComfyUI
InvokeAI: https://github.com/invoke-ai/InvokeAI
Kohya: https://github.com/bmaltais/kohya_ss

For models: you can look at https://civitai.com/models

News:

AI