-
Notifications
You must be signed in to change notification settings - Fork 343
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
RuntimeError: Could not find MSVC/GCC/CLANG installation on this computer. #13
Comments
Hello @justinjohn0306 , conda env create -f environment.yml As it didnt creat the enviroment and didnt install any package. thanks a lot for the attention. About your problem, try this, as stated on https://github.com/NVlabs/stylegan2. I think maybe it is what you need (from my previous attemp on other codes installation) On Windows, the compilation requires Microsoft Visual Studio to be in PATH. We recommend installing Visual Studio Community Edition and adding into PATH using "C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\VC\Auxiliary\Build\vcvars64.bat". |
if you already have Visual Studio installed, just execute this on conda prompt |
Hi there, I had to manually install the dependencies and excluded some of them as they were build for Linux. |
Thanks a lot Justin.... |
yes, and one more thing it looks like you'll have to install Microsoft Visual Studio version between 2013-2017 for this to work. |
Looks like its missing some header file to compile. I would suggest you to look at tensorflow issues about this missing file "op.h" like this issue on another github I cant help better because I'm working now at my job, and cant test it at the moment |
Thanks a lot! |
This could help solve compiling issues on Windows. |
N.B. the stylegan2-ada 2020 project is a version bump from stylegan2 2019 - file formats are compatible. It would take a few hours to switch in the dnn lib - and it fixes glitches found in stylegan2 architecture. |
@johndpope Maybe you should open an issue instead of flooding each issue :) |
RuntimeError: Could not find MSVC/GCC/CLANG installation on this computer. Check compiler_bindir_search_path list in "C:\Users\Administrator\styleflow\dnnlib\tflib\custom_ops.py". I'm having a similar issue. I already did: C:"Program Files (x86)""Microsoft Visual Studio"\2019\Community\VC\Auxiliary\Build\vcvars64.bat Seems to have not helped. Any suggestions? |
Hello, Are you still fighting to install your environment and run the main.py file? (because it is really a fight against the machine lol) I manage to install it on my Windows 10 computer. The command "conda env create -f environment.yml" never worked for me, I had to install each package one by one and some of them failed so I looked for the good channel for the requested version to modify the "-c" argument and relaunch. Some of the packages needed CUDA or CMAKE so I installed them at the end. I could do another post for this. For pytorch I use the following command (after installing cuda 10.0) : conda install pytorch==1.1.0 cudatoolkit=10.0 -c pytorch For pycuda I use this command : pip install --global-option=build_ext --global-option="-IC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\include" --global-option="-LC:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v10.0\lib\x64" pycuda At some point I had the MSVC/GCC/CLANG problem. My issue was that didn't have a "VC" folder in the folder "Microsoft Visual Studio\2017\Community". Check the file "\dnnlib\tflib\custom_ops.py" and edit, if necessary, the value of the variable compiler_bindir_search_path. Be sure that the path is where you have a "cl.exe" file. I also added this path to the PATH of windows, just to be sure. (a reboot might be necessary to be sure that the PATH has been modify properly). Another thing I did is to follow the Requirements of https://github.com/NVlabs/stylegan2 which means I managed to have a clean CUDA install by running nvcc test_nvcc.cu -o test_nvcc -run (you need to clone the git before). After all that you should be good to launch the main.py. (be sure to activate your environment : conda activate <name_of_env>) |
@provostm Yes I got it figured out on windows. Now the only issue I have is the that gui is cut off on the right side and i can't change display size since I'm on an ec2 rdp client |
Can you modify the scripts ? If yes you can follow this issue : ##19 |
@provostm I have CUDA 10.0 installed and as per the StyleGan2 reqs When I attempt to run here is the error https://pastebin.com/x4d2mpyf |
@Machspeedz, your error line 3 says that nvcc is in /usr/local/cuda/bin/nvcc which is wrong. |
Sorry for the confusion, but remove the path from where? I don't remember ever specifying this path. |
@Machspeedz oh right, it is in the dnnlib/tflib/custom_ops.py file at line 65. (you can look the entire code of the repository by searching in github) |
Even after running
I am still getting the error message:
This is what I had in custom_ops.py compiler_bindir_search_path = [ I solved the problem by changing custom_ops.py to: as my Visual Studio was actually: |
I fixed it! I found out that Visual studio 2022 install in program file, not program file x86, so just need to modify custom_ops.py : |
This method is worked, and I have solved it too! |
You mean that I don't need to modify the first four lines? |
I've been agonizing with this problem all night tonight, and I've solved it. Just by trying to figure out the right way to do it. Very much misled by what people wrote about the file |
The text was updated successfully, but these errors were encountered: