QualityScaler
QualityScaler copied to clipboard
Djdefrag
Linux Support
Hey, this project looks interesting and works good on windows, though it would be great, if this could be used on linux aswell. A great usecase for this would be inside a docker container which one can access through vnc or similar. This could aswell be run on a small Linux server ( with a nvidia gpu for cuda), rather than the main machine, without consuming storage and cpu/gpu resources, leaving the main system free for different work (as the process can take quite long for larger files)
Hi, thank you :D
Yes, the project can evolve in many ways, i choosed Windows just because with Pytorch-directml in version 2.0.0 just released it is now compatible with all Gpus :)
Thanks for your reply, but isnt it possible to adjust it, so it could work in linux? Linux (and even docker containers) can use gpu´s and pytorch-directml is, as far as i can see, is available for linux aswell
Btw, a simple py script that could use some arguments, without the gui, would be suitible for use in linux, so in my opinion, there is no need for a gui in linux, which maybe might make it simpler?
@Joly0 GUI works on Linux with minor changes to code. It's likely Pytorch-directml should be replaced by torch on Linux. There are also some windll dependencies that are a problem. Upscaling itself fails for me though.
Wow! Thank you @faveoled! Can you share some info how did you do it? Wine maibe? :D
In version 1.5 i was using Pytorch + cuda, so something can be found there; in fact it only needs to change the code from ".to("dml")" in ".to("cuda")"
@Djdefrag No wine. Tkinter is cross platform, no issue here. Just installed necessary packages and commented out windll calls.
Perfect, maybe i can integrate a if statement for windll calls only if os is Windows, thank you :D
I'm interested in this as well! I tried running on linux with wine, but it doesn't work at all. I would really like to use this app and can't right now!
@Djdefrag Any progress on this one? It would be super usefull, if this could be run on linux
HI @Joly0, no news about this sorry. I recently changed jobs and can't devote much time to these side projects.
To make it work there are two things to do on linux:
- Change from pytorch-directml to pytorch (will work only on nvidia cards and cpu)
- Change input filepaths, for example change the "/" in the paths with os.sep()
I think that all need to make it run on linux. I m really sorry i can't work to the project
Hey, did anything change regarding linux support? I would really appreciate to have this to work in linux
I've found a similar project called Upscayl: https://github.com/upscayl/upscayl Works for me on Linux.
Hi :D,
sadly no news about linux version. The pytorch library used is specifically for Windows to be able to use AI on any gpus :(
I've found a similar project called Upscayl: https://github.com/upscayl/upscayl Works for me on Linux.
I tinkered quite alot with it for docker, but i didnt get it to work, so it isnt really a proper alterantive
Alongside Linux support, at least some CLI support for macOS would be welcome, as well, especially with the newer m-series chips being leveraged effectively for AI processing. There may be some success attempting to push the program as a .exe through WINE but it would be nice to avoid any potential failure points or inefficiencies resulting from that.
Hi @ChristianMGr93,
pytorch-directml, the library used to execute the AI is only compatible with Windows because it uses Directx12 API.
This means that a macOS port is not possible, but I'm working for a Linux port via some sort of translation from Directx12 to Vulkan but with no success at the moment :(