Step1: 60s tutorial ez
For 50 series 5060+ run switch-branch-toole and set to dev by typing "3" in it, then start installing
since it requires pytorch 2.7.0+ cuda 12.8+
Step 2: Download your preferred model(s) from anywhere:I use one called: "
bukkake_20_training_images_2020_max_training_steps_woman_class_word.ckpt"
(>temporary download link<)
put in the folder: stable diffusion -> models -> stable diffusion
If problems: copy paste the error messages to chatGPT
DONE
-------------------------------------------------------------------
How to use:
select your trained ai bukkake model (top left corner)
Go to img2img -> inpainting
Prompt:
brown hair girl gets facialized, (((bukkake))), facial, cum, sticky white cum, large cum amounts, facial
Negative Prompt:
lowres, bad anatomy, bad hands, text, error, missing fingers, extra digit, fewer digits, cropped, worst quality, low quality, normal quality, jpeg artifacts, signature, watermark, username, blurry, anime, drawing, illustration
Mask blur: 5-15 (I prefer 15, but if the face is far away from the camera, go lower)
Mask: Inpaint Masked
Masked Content: original
Inpaint At Full Resolution: Checked (make sure this is check box is checked) <-. It is renamed to Inpaint area and you're supposed to choose "Only masked"
Crop and Resize: ~600x600, increase to make the cum thinner, decrease to make it thicker
Sampling Steps: ~55
Sampling Method: Euler A (others can work great too)
Restore Faces: Off
CFG Scale: ~15 (intensity of the facial)(I sometimes go all the way up to 30)
Denoising strength: ~.2 (how much of the face is changed)
Other notes I would have is I wouldn't mask over the eyes/nose/lips usually, though sometimes it comes out fine, sometimes it comes out borked.
(I sometimes smooth them out afterwards with Photoshop)
-------------------------------------------------------------------------
unnecessary extra stuff:
Install Python 3.10 (Windows 7 ver) (page) https://www.python.org/ftp/python/3.10.6/python-3.10.6-amd64.exe
-Make sure to choose "add to PATH" when installing
(Optional):
This reduces VRAM, and allows you to generate at larger resolutions or batch sizes for a <10% loss in raw generation speed
(For me, singular results were slightly slower, but generating with a batch size of 4 made each result 25% faster on average)
This is recommended for most users
-Edit webui-user.bat
-Change COMMANDLINE_ARGS= to COMMANDLINE_ARGS=--medvram
Launch webui-user.bat, Open it as normal user, not as administrator.
Wait patiently while it installs dependencies and does a first time run.
It may seem "stuck" but it isn't. It may take up to 10-15 minutes.
And you're done!
If no Nvidia it might still work, if not try: Edit webui-user.bat -> set COMMANDLINE_ARGS=--medvram --precision full --no-half --skip-torch-cuda-test
(facialized + corrected)
