search

AlexeyAB / darknet

YOLOv4 / Scaled-YOLOv4 / YOLO - Neural Networks for Object Detection (Windows and Linux version of Darknet )
http://pjreddie.com/darknet/
Other
21.65k stars 7.96k forks source link

Building error #173

Closed FKrjak closed 7 years ago

FKrjak commented 7 years ago

Hi, Sorry for bothering you but I run into some error during the building process. Similarly to this issue https://github.com/AlexeyAB/darknet/issues/136#issuecomment-318465726 but not exactly same, and also your stated fix didn't work for me. After I try to build darknet on windows I get this error:

Severity Code Description Project File Line Suppression State Error MSB3721 The command ""C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\bin\nvcc.exe" -gencode=arch=compute_30,code=\"sm_30,compute_30\" -gencode=arch=compute_52,code=\"sm_52,compute_52\" --use-local-env --cl-version 2015 -ccbin "C:\Program Files (x86)\Microsoft Visual Studio 14.0\VC\bin\x86_amd64" -IC:\opencv_3.0\opencv\build\include -I..\..\3rdparty\include -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\include" -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\include" -I\include -I"C:\Program Files (x86)\Windows Kits\10\Lib\10.0.10240.0\um\x64" -I"C:\Program Files (x86)\Windows Kits\10\Lib\10.0.10240.0\ucrt\x64" -IC:\opencv_2.4.9\opencv\build\x64\vc12\lib -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v9.0lib\x64" -I\lib\x64 -I"C:\Program Files (x86)\Windows Kits\10\Lib\10.0.14393.0\ucrt\x64" -I"C:\Program Files (x86)\Windows Kits\10\Lib\10.0.14393.0\um\x64" -I"C:\Program Files (x86)\Windows Kits\10\Include\10.0.14393.0\ucrt\\" -IC:\opencv_2.4.9\opencv\build\include -I"C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v8.0\include" --keep-dir x64\Release -maxrregcount=0 --machine 64 --compile -cudart static -DCUDNN -DOPENCV -D_TIMESPEC_DEFINED -D_CRT_SECURE_NO_WARNINGS -D_CRT_RAND_S -DGPU -DWIN32 -DNDEBUG -D_CONSOLE -D_LIB -D_MBCS -Xcompiler "/EHsc /W3 /nologo /O2 /FS /Zi /MD " -o x64\Release\dropout_layer_kernels.cu.obj "C:\Users\fredy\Desktop\darknet-master (1)\darknet-master\src\dropout_layer_kernels.cu"" exited with code 1. darknet C:\Program Files (x86)\MSBuild\Microsoft.Cpp\v4.0\V140\BuildCustomizations\CUDA 8.0.targets 689 11 times. I tried everything available but I couldn't find the answer. Even tried newer version of CUDA but had same error. If you have any ideas how to fix the issue, I would be thankful.

Thank you for your answer FKKrjakonzo

MyVanitar commented 7 years ago

I have compiled the Alex repo in counting with my hairs, so I'm sure there is no problem with the Build section. if all components are in place, you will not face any error. I suggest that you install the CUDA again, but before that, remove all traces from add/remove programs and also all traces from from your hard disk. it saves some files in several places. Also re-open the Visual Studio if you change the CuDNN path in the environmental variables, otherwise it reads the previous path.

FKrjak commented 7 years ago

May I ask you, how did you manage to put all components in place? I installed VS2015, then installed Cuda 8.0, after that I copied openCV3.0 and put links to PC variables in paths "C:\opencv_3.0\opencv\build\include" and "C:\opencv_3.0\opencv\build\x64\vc14\lib". Then I started darknet in VS and set release and x64. I think I did everything as said. If you can suggest anything more I have to do please do so.

MyVanitar commented 7 years ago

make sure the paths to your OpenCV are correct. Besides make sure that you have modified the OpenCV parameters to match with your installed OpenCV version. I mean inside detector.c

as the Alexey mentioned, check the paths inside pre-processors and Linker of your project properties.

FKrjak commented 7 years ago

All paths in detector.c seems to be working fine ( https://snag.gy/rdqeEf.jpg ), also I am using CV 3.0 so by default it should be set. Also, I checked everything from https://snag.gy/AlUFVY.jpg. I am kind of desperate with it :/ .

MyVanitar commented 7 years ago

I'm sure there is something wrong with your definitions, because now I'm working with it also and compiling it almost every minutes :-)

MyVanitar commented 7 years ago

try to run this command in your CMD and see what gives you: nvcc --version

FKrjak commented 7 years ago

nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2016 NVIDIA Corporation Built on Mon_Jan__9_17:32:33_CST_2017 Cuda compilation tools, release 8.0, V8.0.60

MyVanitar commented 7 years ago

that's correct but again I believe something wrong with paths. I can guarantee you the repo compiles correctly.

FKrjak commented 7 years ago

Yeah, I agree but I can't find what. I double checked everything and I am out of ideas, everyone managed to build darknet easily and I am struggling hard to get it work :/ If you can find anything I would be glad, anyway thank you very much for trying https://snag.gy/pgUSui.jpg

AlexeyAB commented 7 years ago

@FKKrjakonzo

AlexeyAB commented 7 years ago

@FKKrjakonzo Also show path, where is file opencv_world320.lib located in your disk? And if you have another file, then change this line, for example opencv_world300.lib: https://github.com/AlexeyAB/darknet/blob/0316685077bf82d4dbeadb9e947e3ea16b6ae703/src/detector.c#L16

FKrjak commented 7 years ago

I reainstalled cuda several times but I can't even build CUDA sample :/

AlexeyAB commented 7 years ago

@FKKrjakonzo So your problem doesn't relate to Darknet, your error relates to CUDA-installation in general: https://www.google.ru/search?q=MSB3721+CUDA

Also check PATH variable (check paths to MSVS, NVIDIA and others) in the Windows: Control Panel -> System -> Advanced System Settings -> Advanced -> Environment Variable

FKrjak commented 7 years ago

Sorry for bothering you guys, and thank you very much for every advice. I just tried installing darknet on another PC, and it worked just right with no problems. Now I have to find the way to run it on my PC.