Hi,
I tackle with a lot of difficulties to compile Pytorch libraries on my Windows 10, it is a PC with a NVIDIA QUADRO K4200 (with compute capability. 3.0). The CUDA driver on the graphic board is CUDA 11.4.94 but the runtime library library is 10.2.89. I have installed VS 2019 with MSVC v140-VS 2015 c++ build tools (v14.00), MSVC v142-VS 2019 (v14.27) and (v14.28-16.8), also I installed MSVC v140 - VS 2015 C++ build tools (v14.00) and last Windows 10 SDK. Honestly, I don’t know if any of these VS libraries are redundant or not for the compilation of the pytorch. After 6 hours of compilation (set CMAKE_GENERATOR=Visual Studio 15 2017) I installed pytorch library on python and I see follwing error:
c:\Users\PC\python
import torch
Traceback (most recent call last):
File “”, line 1, in
File “C:\Users\PC\anaconda3\envs\rnnenv\lib\site-packages\torch_init_.py”, line 125, in
raise err
OSError: [WinError 1114] A dynamic link library (DLL) initialization routine failed. Error loading “C:\Users\PC\anaconda3\envs\rnnenv\lib\site-packages\torch\lib\backend_with_compiler.dll” or one of its dependencies.
Other print-out is the following:
import ctypes
ctype.CDLL(r"C:\Users\PC\vcomp140.dll")
Traceback (most recent call last):
File “”, line 1, in
NameError: name ‘ctype’ is not defined
ctypes.CDLL(r"C:\Users\PC\vcomp140.dll")
Traceback (most recent call last):
File “”, line 1, in
File “C:\Users\PC\anaconda3\envs\rnnenv\lib\ctypes_init_.py”, line 382, in init
self._handle = _dlopen(self._name, mode)
FileNotFoundError: Could not find module ‘C:\Users\PC\vcomp140.dll’ (or one of its dependencies). Try using the full path with constructor syntax.
ctypes.CDLL(r"C:\Users\PC\anaconda3\vcomp140.dll")
<CDLL ‘C:\Users\PC\anaconda3\vcomp140.dll’, handle 7ff8a7ec0000 at 0x26561a54eb0>
Please, any idea before starting again all the installation process?
Br/Luis