-
Notifications
You must be signed in to change notification settings - Fork 3.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Trying to compile DLIB with CUDNN. CUDA v11 with CUDNN v8 Windows 10, Python #2108
Comments
I am having the same problem on ubuntu. I think the current repo is not yet supporting these newer versions of CUDA and cuDNN. |
This might be the case however I struggled to find cases of this among other people so I can only imagine it isn't as widespread of an issue. Furthermore I tried an earlier version as I had mentioned which also led to errors |
I just updated dlib to support cuDNN 8.0. So try the latest code on github. |
Just tried downloading from github and installing it with pip. Ran into the exact same issue and error as quoted |
What did you type exactly? |
I downloaded dlib from github then did pip install -v D:\dlib-master |
Go into the dlib folder and type this instead and paste the output:
|
Hey, I managed to figure out how to do the build thing. Here's the output from that. (venv) D:\dlib-master\dlib>cmake .. |
That test is just trying to build the
What does that say? |
Actually, just pull the latest dlib from github and try building again. I updated the cmake settings so you should get more informative output. So just pull from github and do the cmake build again, and post the output here. |
Okay so this is the error I get. Seems like a file is missing but I'm not sure which since I installed cudnn properly to the best of my knowledge. If you could tell me how to proceed I'd be highly grateful.
:VisualStudioVersion=16.0 /v:m && Microsoft (R) Build Engine version 16.6.0+5ff7b0c9e for .NET Framework |
I would download cuDNN again and make sure you get that missing file. |
Okay so it turns out CUDNN v8 has some extra files in the include folder that you have to copy into the CUDA folder along with the normal 3 files. That seems to have resolved it. Thank you for the help |
For anyone else having issues with Re-run the dlib installation, but ensure you don't use the cache: You should see the following during installation:
|
Which include folder are you referring to? Thanks!
|
Thank you so much! It really helped! I manually downloaded source code and tried to compile dlib with cmake and after trying every way I still couldn't make it. This is simple and helpful. |
##Version Details
Windows 10
Cmake 3.17.3
dlib 19.20.0
CUDA v11
CUDNN v8
python 3.8.3
NVIDIA Geforce 1050 TI
##Summary
I am trying to do facial recognition and have reached a roadblock. I am unable to compile DLIB with CUDNN as it gives me an error saying CUDNN is found but is the wrong version. I tried installing CUDA v10.2 with CUDNN v7.6.5 and that seemed to compile with CUDNN however I ran into an issue where I get the following error: ImportError: DLL load failed: The specified module could not be found
Expected Behavior
DLIB Compiles with CUDNN
Current Behavior
-- Found CUDA: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.0 (found suitable version "11.0", minimum required is "7.5")
-- Looking for cuDNN install...
-- Found cuDNN: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v11.0/lib/x64/cudnn.lib
-- Building a CUDA test project to see if your compiler is compatible with CUDA...
-- Checking if you have the right version of cuDNN installed.
-- *** Found cuDNN, but it looks like the wrong version so dlib will not use it. ***
-- *** Dlib requires cuDNN V5.0 OR GREATER. Since cuDNN is not found DLIB WILL NOT USE CUDA. ***
-- *** If you have cuDNN then set CMAKE_PREFIX_PATH to include cuDNN's folder. ***
-- Disabling CUDA support for dlib. DLIB WILL NOT USE CUDA
-- C++11 activated.
##Request
I am completely new to programming so I would be very grateful if you could be clear as to what steps I need to take and how to take them.
The text was updated successfully, but these errors were encountered: