I’ve recently started to use Azure and installed the
NVIDIA GPU-Optimized Image for PyTorch release
21.11.0 on a
NCv3 machine and I have a few questions:
First, when logging in, I get the message:
Welcome to the NVIDIA GPU Cloud image for TensorFlow.
Is the image correct? I triple checked that I installed the pytorch image. How do I check the image version when logged into the VM itself?
Also, when importing pytorch in the container in the python interpreter I get a module not found error:
Python 3.8.10 (default, Mar 15 2022, 12:22:08) [GCC 9.4.0] on linux Type "help", "copyright", "credits" or "license" for more information. >>> import torch Traceback (most recent call last): File "<stdin>", line 1, in <module> ModuleNotFoundError: No module named 'torch' >>>
To summarise, how do I use pytorch in this image? Just pip install? If so, what pytorch +CUDA version, because I’m not sure anymore if the documentation on NGC applies to Azure images