Photos & Videos AI Tools: How to and Technical Discussion on Creating NSFW Gay Sex Stuff with Stable Diffusion (and others)
- Thread starter StudmeatFL
- Start date
-
- Tags
- bara civitai men stable diffusion
idk if this isn’t the group for this (just lmk and ill delete this message) but like be fr we’re all horny, if ur into making super realistic ai generated (unrestricted prompts) models this site is so good. Male support is the best I’ve seen DeepMode | AI Character Creation & Art Generator
No, not exactly. In my case I’m not trying to implement inpainting. I’m simply attempting to use text to image, using images uploaded via ControlNet as a guide (usually Canny and/or Depth Midas). I did notice today when launching the Automatic1111 GUI a ControlNet error message saying “insightface” not installed, try installing manually, etc. But I apparently don’t know enough about what I’m doing to remedy this, because I tried ChatGPT and Googling a fix and still came up short.
That sounds like you’re missing one of the Control Net libraries. I’m not as familiar with troubleshooting installs, either.
InsightFace is a python library:
pip install insightface and https://insightface.ai · Issue #891 · deepinsight/insightface
It's fine advertising these. There are a handful of online cloud based private servers/services to host your creations. This one reminds me a little bit of DeepSwap (Deepswap - Best AI Face Swap Online for Video & Photo) in that it seems more specifically tailored to generating your own private 'models'.
Yeah, this all feels a bit over my head right now as a casual user/hobbyist. For now, I’ve scrapped my entire SD install and plan to start over from scratch, see where that goes. From what I can gather, Automatic1111 is still considered by many to be a legitimate and feasible GUI, although there are those who argue Forge or Comfy are superior. I imagine, to some extent it probably depends on your preferred work flow, experience level and objectives.
Yeah I don't have the hardware to run anything like that locally. I was using Google Collab and one of the notebooks, but I would run into issues like you're discussing where I'd have to re-run python libraries here and there. Too much work.
I did drop into Tensor.art for a while, but I've noticed they're not consistent with their libraries and lora, so I'm still looking for a new 'home' for my art projects. Life kinda takes over and you have to focus on that for a while instead.
I hear that. I’m getting quite decent generation times now, but the troubleshooting and fine-tuning still eats up my free time fast. I did give Forge a go last night — at least with SD; didn’t try Fusion or the other models yet. It was a cinch to install, and while the generation speed is indeed a bit faster, I still couldn’t get it to work properly with ControlNet. Then I switched back to Automatic (fresh install) and experimented some more. The bizarre thing is, sometimes ControlNet seems to work fine and then sometimes it just doesn’t at all — for no apparent reason. It would be interesting to see how it runs on different hardware, because at this point I’m really just grasping at straws. I may try the Collab route, just to compare results with the same parameters, as an experiment.
I really liked Google Collab until it would blow up and I’d have to rebuild. The nice thing is when using it in that mode, you can point to another directory for models, LORA, etc so you only have to reinstall and repoint.
I’ve had the odd issues with Control Net. I could honestly only get it to work properly with Depth; Open Pose randomly worked, most things didn’t and I blamed it on “user error”. I’ve always had much better results using a combination of ControlNet Depth with Img2Img. Usually in these situations, it was much like my former example where I took a base image I like for the pose and pasted in some additional cocks coming in off frame to get the orgy feel I was looking for. As long as the images you’re using are fairly clear along with anything you’re pasting in, SD does a fairly good job with interpretation.
Good luck!
Yep, just wanted to make sure they weren't coming into the technical discussion
You can also do that using the DeepSwap.ai I linked before: Deepswap - Best AI Face Swap Online for Video & Photo
Most people will not do face-swapping with real celebrities, but you can find some AI artists that offer those services by paying into their Patreon, etc.
I
I was wondering if there is a checkpoint/ sd model which can generate realistic fit men with the body type in the images below because with virile stallion v5 i always get beefy men (which i love)
I have tried it and love the images i get!xpost from the previous thread - how to use SD without a PC/laptop
sign up here: https://tensor.art/images/712535220...0710122895&source_id=nj-0pl_hlEG0rfosb3f28RUg
go here: Virile Stallion - V5.0 Photo | Stable Diffusion Model - Checkpoint
some setting:
Press |> Run;
The author recommends the following parameters. Do you want to apply them? - press Use Recommended
Prompt:
score_9, score_8_up, score_7_up,
cinematic, realistic, 1boy, muscular male
Negative prompt:
low quality, lowres, bad anatomy, normal quality, worst quality, 3D, render, illustration, drawing, comic, watermark, greyscale, monochrome
Follow these settings, have fun!
![1.png]( "1.png")
![2.png]( "2.png")
![3.png]( "3.png")
it'll take time to load and u shud be able to see ur creation there
![4.png]( "4.png")
I was wondering if there is a checkpoint/ sd model which can generate realistic fit men with the body type in the images below because with virile stallion v5 i always get beefy men (which i love)
Attachments
If anyone could help, I'm having a lot of issues using LORAs of people's likenesses (I don't plan on uploading this) wherein it never actually ends up looking like them. I'm using Forge if that helps.