1 min readMay 8, 2020
Well, installing PyTorch locally to support GPU training/scoring is straightforward following PyTorch website guidance.
If you mean running the same steps in the post in a container with a GPU, then this needs a bit of extra work.
Unfortunately, the GPU containers on Azure support CUDA 9.0 which is not enough for PyTorch as of now. Other cloud providers may have a solution.