It was caused by the missing of some of the dependencies: mkl, mkl-fft, intel-openmp and VC 2017 Redist.
For conda packages:
conda install mkl mkl-fft intel-openmp numpy
# If package not found, do this
conda update conda
For wheel packages:
pip install mkl mkl-fft intel-openmp numpy
And then download VC 2017 Redist and install it.
A reboot may be required for the libraries to take effect.
If this persists, it may be related to your environmental variables, such as PATH
.
Make sure it is not broken and at least point to the following locations:
[PythonRoot]
[PythonRoot]\Library\bin
[PythonRoot]\Scripts
[PythonRoot]\bin
%SYSTEMROOT%
P.S:
#7579 ImportError: DLL load failed: The operating system cannot run %1.
does not seem to be the same issue with this.
It is related to Anaconda. Fix command: conda install -c defaults intel-openmp -f
.
If you want to help us, please provide us the logs with python -v -c 'import torch'
and the result of the following code:
# pip install psutil first
import psutil, os
p = psutil.Process( os.getpid() )
for dll in p.memory_maps():
print(dll.path)
After a long search i found the solution
The problem is that if the GPU is old the pytorch version before 0.4.0 doesn’t work, so you have to install the newest pytorch,
run this - conda install -c pytorch pytorch
After you install the newest pytorch you will face torch._C import * DLL load failed problem
to fix that run the ff code
set PYTORCH_BUILD_VERSION=0.4.1