Open billpage opened 6 years ago
Work-round:
import os
os.environ["MKL_THREADING_LAYER"] = "GNU"
Restart kernel.
I could not verify the problem with the latest version. My output is:
Using cuDNN version 7005 on context None
Mapped name None to device cuda0: GeForce GTX 780 (0000:02:00.0)
I could not verify the problem with the latest version. My output is:
Using cuDNN version 7005 on context None Mapped name None to device cuda0: GeForce GTX 780 (0000:02:00.0)
If someone else came to here from Google search. One solution could be defining cuda first.
[global]
device = cuda
floatX = float32
With Jyptyer worksheet anaconda_kernel:
I get the following error message: