Torch.cuda.is_available() returns false after suspend

2

You could try to reload the nvidia kernel module as described here.

1 Like