search

huchenlei / stable-diffusion-ps-pea

Use Stable Diffusion in Photopea!
GNU General Public License v3.0
74 stars 3 forks source link

Adding this repo to the SD Auto1111 #3

Closed p0mad closed 1 year ago

p0mad commented 1 year ago

Is there an existing issue for this?

What happened?

Hi, I want to do exactly the same process as the image https://user-images.githubusercontent.com/20929282/256951742-5dcb6d6f-5c3e-4cf8-abf6-c5223059a8af.png into the auto1111 but without using external applications like photoshop /pea.

Actually i want to concatenate controlnet Pose ( openpose as a unit0 of ControlNet) and CN canny (canny as a unit1 of Controlnet) but not just the whole canny image, just the hands part!

is it possible to use this repo? any guide or insight in that regards would be much appritiated

Thnaks Best regards

Steps to reproduce the problem

  1. Go to Auto1111, ControlNet tabs and add openpose and canny Model
  2. edit canny Model to just include hand parts!
  3. Auto detect hands to be used in canny CN unit1 for video generation

What should have happened?

autodetect hands and just use those parts for canny!

Commit where the problem happens

webui: stable-diffusion-ps-pea: [Previous good version]

What browsers do you use to access the UI ?

No response

Command Line Arguments

no

Console logs

no

Additional information

No response

huchenlei commented 1 year ago

Currently you can use https://github.com/yankooliveira/sd-webui-photopea-embed to send preprocessor maps to ControlNet (You can specify which unit to send to).

I think autodetection of hand is out of current scope of the extension here, as the extension does not have capability of invoking an hand detection algorithm in the browser. There need to be some sort of backend implementation in A1111 to do that first.