Hi,
In the pytorch installation documentation, it is mentioned to check the compatibility matrix:
to install the correct pytorch version but this matrix is inconsistent with what I can download in the Jetson Download Center | NVIDIA Developer
I cannot find the compatible pytorch wheels for the JetPack 6.1 (2.6.0a0+df5bbc0, 2.5.0a0+e000cf0ad9, 2.5.0a0+b465a5843b). Where are they ? Do I have to build pytorch from source for the JetPack 6.1?
PS: I tried to install the only pytorch version that I can download for JetPack 6.1 and the wheel is not supported
Hi,
Please find the container in the below link:
Thanks.
Thanks a lot @AastaLLL , you’re suggesting to use a docker container in the jetson, but is this the only way ?
I thought I could create a simple virtual environment directly in the Jetson OS (Ubuntu 22.04) and install pytorch inside. Am I wrong ?
I was finally able to install pytorch in JetPack 6.1 after @AastaLLL 's answer on this topic.
My guess is that the compatibility matrix may not be up-to-date and that one who wants to install pytorch on JetPack 6.1 in a virtual environment needs to use the following wheel:
# activate your venv with python 3.10 then
pip install --no-cache-dir https://developer.download.nvidia.com/compute/redist/jp/v61/pytorch/2.5.0a0+872d972e41
python 3.10 is mandatory because the wheel has been built with it