AMD GPU + Windows + ComfyUI! How to get comfyUI running on windows with an AMD GPU!

Video Statistics and Information

Video
Captions Word Cloud
Reddit Comments
Captions
hey there hom labbers and Engineers Fe engineer here so today we're going to go and take a look at setting up comfy UI in Windows on AMD cards now not that long ago I don't think this was possible so let's take a look at how to get this all set up and running running comfy UI on Windows is very similar to running other versions of stable diffusion including automatic 11 11 direct ml so you will need to have get for Windows and miniconda already set up and running and with miniconda make sure that you add the directory to path we're going to create a new Anaconda environment named comfy UI and we're going to Target Python 31012 and after a minute or two this should be all set up for us our environment now seems to all be set up so we're going to cond activate comfy UI we'll come over to the comfy UI GitHub page go up here find this clone https copy this and then we're going to type get clone paste in the URL and it will clone it all over for us now we should be able to CD into comfy UI and from here we need to pip install torch DD directml once torch direct ml has finished installing after that it's a pretty simple pip install minus r requirements and from there we should be able to just type python main.py D- direct ML and with that if we just go to the URL that it indicates we'll see that comfy UI is in fact up and running and I just drop in a random model into the folder and I just left the default prompts and everything and I just Chang the CFG down to 7.0 and I bumped the steps up to 100 mainly so that people can see the an idea of the performance I am running this on a 7900 XTX but if we just hit Q prompt you'll see I end up getting about six iterations per second which while not necessarily the fastest in the world and I definitely get better when I'm using onnx and Olive and things like that at least with this I can just in Windows run a normal model that is not onnx or anything like that I don't have to convert it I don't have to do anything I can download Straight off of civetti or hugging face plug it in and I am good to go if you have other models available somewhere else like potentially in SD test and stable diffusion direct ml you can grab your models and we are going to use stable diffusion models not onnx Olive models and inside of here I have one called animerge you can simply copy or move we'll just go right back get comfy UI models go to the checkpoints drop the file there now that we have a new model in there we're going to go over here hit refresh and when we hit refresh you'll see that this other animerge model is now in I can hit get Q prompt it will render away and comfyi seems to be up and running for any folks who are wanting to create a batch file to run this automatically I haven't exactly gotten the cleanest version of it but effectively run comfy UI dobat put this inside of it call cond to activate comfy UI CD comfy UI Y and then your script python main. py-- directml for the moment for me at least I must open an andaconda mini prompt so right inside of here and then all I have to do is run- C run comfy UI dobat and this does in fact work and everything loads up and everything is fine I cannot quite yet figure out a good way to do it so that you can either double click on the file or run it from command prompt but this is a bit of a shortcut for those who want it to be a little bit easier and don't want to have to remember quite as many commands I know this video is pretty short but I really wanted to give people the ability to run comfy UI on AMD graphics cards in Windows and I didn't didn't want to add in a whole bunch of fluff about this is how you use comfy UI and this is what you can do with it and all these things I think people are generally pretty aware of comfy UI it is very complicated to use but along with all of that complication it is very powerful in the things that it can do there are tons and tons and tons of YouTubers out there who go in-depth into setting up come vui to do all sorts of crazy things but with this people should be able to get comfy UI up and running on their AMD graphics cards and with that you should have some interesting and fun things to do in the new year thanks so much everybody thank you so much for watching my videos hom labbers and Engineers I create and edit all these videos on my own so any likes and subscribes will massively help out the channel and allow me to continue creating content to help people if you got value out of this please please consider liking and subscribing to the channel to be notified when new content drops if there's something I've not covered but you would like to see a video on it please leave a comment down below and again a massive thank you to everyone I hope you have a great day
Info
Channel: FE-Engineer
Views: 5,597
Rating: undefined out of 5
Keywords: Stable Diffusion, DirectML, Windows, AMD GPU, Amd stable diffusion, ComfyUI
Id: 8rB7RqKvU5U
Channel Id: undefined
Length: 7min 15sec (435 seconds)
Published: Sun Dec 24 2023
Related Videos
Note
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.