I have CUDA 9.0 installed on my Ubuntu 18.04, and it success to compile on GCC.
however, when I opened a project with the pycharm 2016.3 (python 3.67), the checking line “torch.cuda.is_available()” return False.
do you have an idea how to fix it? maybe I have to define some configuration on the pycharm before i used the cuda?
thanks!