Forcing CPU usage when a compatible GPU is present

Part 1 (2019)
2

i had the same problem
i got it to work o fastai2 with
fastai.torch_core.default_device(use_cuda=False)

1 Like