Closed lovettchris closed 1 year ago
hi @lovettchris, thanks for adding this. I will approve the change, but please be aware that we will probably update these samples at some point to use the DirectML backend for PyTorch (which is a work in progress right now).
The sample was broken when it chose to use my CUDA GPU (the testAccuracy function forgot to move the images to the device and the prediction from the device).