#Problem with ONNX ML Agents
1 messages · Page 1 of 1 (latest)
yeah
It seems like this was an issue that was fixed, but not included in a release yet.
https://github.com/Unity-Technologies/ml-agents/commit/631901b53cd75f6b9e03e3cd1fd6279047ed0375
So i clone this using the command?
I think you can downgrade pytorch to 2.1.1 and it should work.
You could clone the github repo, but that would probably bring in a lot of stuff that you don't really need in your project.
And there might be some issues due to ml-agent source code colliding with the pip installed version.
so downgrade pytorch to 2.1.1 only?
Yes.
If that doesn't help, you should share more details. That screenshot you shared omits way too much context and details.
alright
whats the command for downgrading again
Is this correct?
from pytorch website
my cuda version is 13.0
I installed the CUDA 12.1
yoo
it works now
tysm dlich
I had this issue for some time now
heyy @ocean granite I switched pcs and now I encountered another problem
Well, you basically need to remove the packages which versions don't match the requirements and reinstall the correct one.
pip3 uninstall onnx && pip3 install onnx=1.15.0
is the command like this?
I think so. I don't really memorize pip commands. Check the docs, google or AI chatbot.