TheFerumn

If you are using process lasso or some similar programs then its possible that compilation process is working on 1 core of your processor only. Then it will work forever....

What version of python you installed ?

Probably wrong rocblas libs. Try to download this [rocblas.for.gfx906.7z](https://github.com/likelovewant/ROCmLibs-for-gfx1103-AMD780M-APU/releases/download/v0.5.7/rocblas.for.gfx906.7z) You have to paste it into C:\Program Files\AMD\ROCm\5.7\bin If it won't help try different versions from here: https://github.com/likelovewant/ROCmLibs-for-gfx1103-AMD780M-APU/releases

Hires.fix is using upscalling so maybe you need to change Tile size for the upscaler you use. Check which part of generating is actualy slow. Is it upscaling or sampling...

Full log please and what exactly you tried to generate ? which type of model you tried etc.

I guess you tried already clean install ? I am using Forge DirectML with Reactor extension without any big issues. Info about ControlNet Preprocessor location shows always during launch so...

Can you show what errors do you have when trying other faceswap extensions ? Also check if your Reactor scripts folder looks like this and it correctly downloaded everything for...

I have no idea why it happens to you IF its clean installation. It has nothing to do with ControlNet so don't even bother looking. The only thing that come...

This is wierd as I was pretty sure script loading is exacly the same no matter if you use Zluda or DML but it doesn't look like it and since...

If its not a big deal and i saw no one else posting this issue i will suggest you to completely uninstall whole python and make sure to delete all...