refinenet-pytorch
refinenet-pytorch copied to clipboard
run example on CPU
First of all thanks for the code. I'm just starting to play around with these things and find very useful. Just one thing.. to run on CPU I had to add this parameter map_location='cpu' to function load_state_dict of resnet.py, otherwise the notebook will give error on deserialization.(cell 5) and also removed ".cuda()" from here:
segm = mnet.cuda()(img_inp)[0].data.cpu().numpy().transpose(1, 2, 0)
Is this correct or I miss something ?
I want it runs on cpu using on Windows 10, and find the similar problems as @roccopietrini described above