So when I run cuda.select_device(0)
and then cuda.close()
. Pytorch cannot access the GPU again, I know that there is way so that PyTorch can utilize the GPU again without having to restart the kernel. But I forgot how. Does anyone else know?
from numba import cuda
device = cuda.get_current_device()
device.reset()
And then trying to run cuda-based pytorch code yields:
RuntimeError: CUDA error: invalid argument
from Make GPU available again after numba.cuda.close()?
No comments:
Post a Comment