-
-
Notifications
You must be signed in to change notification settings - Fork 60
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Can't select CUDA on Windows (torch with CUDA support installation problems) #192
Comments
Hello There's a bug in the tool's update system which uninstalls torch cuda while performing the update. You should be able to reinstall it by following the instructions here: https://github.com/octimot/StoryToolkitAI/blob/main/INSTALLATION.md#cuda-support BTW, it doesn't look like you're using a virtual environment, are you? This is not optimal since any update on the main Python installation might update the packages StoryToolkitAI needs to run. If possible, I highly recommend setting up a virtual environment: https://github.com/octimot/StoryToolkitAI/blob/main/INSTALLATION.md#5-set-up-a-virtual-environment Cheers! |
Hi, thanks for response, my main installation is inside a virtual environment, this was just a quick test. I've tried reinstalling torch as per the instructions many times at this point, but it still doesn't seem to work. E:>cd E:\StoryToolkitAI E:\StoryToolkitAI>venv\scripts\activate.bat (venv) E:\StoryToolkitAI>py -3.10 -m pip uninstall torch torchaudio torchvision (venv) E:\StoryToolkitAI>py -3.10 -m pip cache purge (venv) E:\StoryToolkitAI>py -3.10 -m pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu118 (venv) E:\StoryToolkitAI>py -3.10 -m venv\Scripts\python StoryToolkitAI\storytoolkitai (venv) E:\StoryToolkitAI>venv\Scripts\python StoryToolkitAI\storytoolkitai |
I think this might occur because you're still using the Instead of:
Simply do:
One other note: once you set everything up in the virtual environment, you don't have to activate it to run the tool, if you're running the command |
Thanks a lot for the help, unfortunately I still can't get it to work. Sorry for the long spam, but here is the log from me deleting everything and trying to do a fresh install: Microsoft Windows [Version 10.0.22631.4037] C:\Users\Alex>e: E:>cd E:\StoryToolkitAI E:\StoryToolkitAI>git clone https://github.com/octimot/StoryToolkitAI.git E:\StoryToolkitAI>py -3.10 -m virtualenv venv E:\StoryToolkitAI>venv\Scripts\activate.bat (venv) E:\StoryToolkitAI>pip install -r StoryToolkitAI\requirements.txt [notice] A new release of pip is available: 24.1.2 -> 24.2 (venv) E:\StoryToolkitAI>pip uninstall torch torchaudio torchvision (venv) E:\StoryToolkitAI>pip cache purge (venv) E:\StoryToolkitAI>pip install torch torchvision torchaudio --extra-index-url https://download.pytorch.org/whl/cu118 [notice] A new release of pip is available: 24.1.2 -> 24.2 (venv) E:\StoryToolkitAI>venv\Scripts\python StoryToolkitAI\storytoolkitai × python setup.py egg_info did not run successfully. note: This error originates from a subprocess, and is likely not a problem with pip. × Encountered error while generating package metadata. note: This is an issue with the package mentioned above, not pip. [notice] A new release of pip is available: 24.1.2 -> 24.2 pip install -r E:\StoryToolkitAI\StoryToolkitAI\requirements.txt Traceback (most recent call last): |
This seems to hit the same issue as it does here: #184 I'll try to recreate the problem on one of our machines as soon as its free and report back with a solution! |
Could you try the following, please:
Cheers! (edit: correction to install torch cuda in the right environment) |
I just pushed d419309 which hopefully patches the installation issue. So, please try a fresh pull and install. Cheers! |
Hi, sorry for the late reply, I'm a bit off grid atm. This is the error I got when trying to run storytoolkit after running a fresh pull and install, the tool won't run at all now. (venv) E:\StoryToolkitAI>venv\Scripts\python StoryToolkitAI\storytoolkitai (venv) E:\StoryToolkitAI> |
This looks like something is messed with your virtual environment. I recommend deleting the venv folder completely and re-creating the virtual environment. |
I've just tried that and it still isn't working, throwing the same error. I also tried reinstalling python (venv) C:\Program Files\StorytoolkitAI>venv\Scripts\python StoryToolkitAI\storytoolkitai (venv) C:\Program Files\StorytoolkitAI> |
On 0.24.2, I installed the cuda version of torch as instructed, but I can't select the gpu for processing, it is running it CPU-bound. Any input welcome!
Edit: I've done a fresh install of storytoolkit, and it is attempting to reinstall torch when I launch it.
As you can see the cuda 2.4.0+cu118 version of torch is installed:
When I launch storytoolkit, it uninstalls that version of torch and reinstalls 1.2.1 without cuda, can you let me know what is going on here, thanks.
The text was updated successfully, but these errors were encountered: