Dlib ImportError in Windows 10 on line _dlib_pybind11 import *, DLL Load Failed

Question

I am able to successfully install Dlib with CUDA support in Windows 10 but getting an error during "import dlib" in my python code of computer vision project.

Environment: Windows 10, Python 3.7.6 (Anaconda), CUDA 11, CuDNN 10.2

Error Message:

>>> import dlib
Traceback (most recent call last):
  File "", line 1, in 
  File "C:\ProgramData\Anaconda3\lib\site-packages\dlib-19.20.99-py3.7-win-amd64.egg\dlib\__init__.py", line 12, in 
    from _dlib_pybind11 import *
ImportError: DLL load failed: The specified module could not be found.

user_13755243 · Answer

This can be solved by copying the cudnn64_7.dll (available here: https://developer.nvidia.com/cudnn) into the %CUDA_PATH%/bin directory (probably something like this: C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.1\bin)

user_6424483 · Answer

I am able to find and fix this issue. CUDA 11 installation wasn't able to add few of the directories into the PATH environment variable (Windows 10). It was truncated due to the max length of 2048 characters. I have removed a few of the unused software paths from PATH value and after reinstallation, dlib 19.20 is working with CUDA 11 now.

I created an issue on DLIB Github under the following link which has more information regarding error logs and snapshots for this issue. https://github.com/davisking/dlib/issues/2097

user_9723722 · Answer

In my environment, the problem was due to an error somewhere in the build process that resulted code to load CuDNN dynamic libraries not being included in the generated file dlib/__init__.py despite having no build error. In my case the file always included this strange block of code:

if 'OFF' == 'ON':
    add_lib_to_dll_path('cudnn-NOTFOUND')
    add_lib_to_dll_path('C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.5/lib/x64/cudart.lib')

The second line's cudnn-NOTFOUND gave a clue of what happened with my build.

As I followed instructions on this page, copying all the binaries and include files to the right places within the CUDA directory, I only needed to modify the code to (similar to what Epic Chen's answer suggests but I got rid of the if clause and the bad code line):

add_lib_to_dll_path('C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.5/lib/x64/cudart.lib')

That workaround fixed the problem for me.

My configuration is CUDA 11.5, CuDNN 8.3.1.22, dlib compiled using Visual Studio 2019. The environment variable CMAKE_PREFIX_PATH to the CuDNN directory to get the compiler to find the include files and libraries.