You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've been reading through this thread opencv/opencv#10963 and seeing lots of discussion #500 and #414 but I can't compile with OpenCV 3.4.0, I get the errors:
1>------ Build started: Project: darknet, Configuration: Debug x64 ------ 1>http_stream.cpp 1>c:\users\winbase\code\opencv\build\include\opencv2\flann\matrix.h(70): warning C4003: not enough actual parameters for macro 'free' 1>c:\users\winbase\code\opencv\build\include\opencv2\flann\matrix.h(70): error C2059: syntax error: ',' 1>c:\users\winbase\code\opencv\build\include\opencv2\flann\matrix.h(87): note: see reference to class template instantiation 'cvflann::Matrix<T>' being compiled 1>c:\users\winbase\code\opencv\build\include\opencv2\core\cuda.hpp(106): error C2059: syntax error: 'constant'
from matrix.h
/** * Convenience function for deallocating the storage data. */ CV_DEPRECATED void free() { fprintf(stderr, "The cvflann::Matrix<T>::free() method is deprecated " "and it does not do any memory deallocation any more. You are" "responsible for deallocating the matrix memory (by doing" "'delete[] matrix.data' for example)"); }
Is there any guidance on which version of OpenCV 3 is compatible with Darknet? I can compile with 2.4.13 but in running darknet detector train it always crashes on try to allocate workspace. Without OpenCV I can train without problems. I'm not sure how to get the OpenCV compatibility. I'm using Windows 10, Visual Studio 2017, CUDA 9.2, and the following flags:
@AlexeyAB Hello! Can you explain me please. For what it answers here // get_region_boxes (l, 1, 1, thresh, probs, boxes, 0, 0); and why is now commented on? Not used or replaced?
I've been reading through this thread opencv/opencv#10963 and seeing lots of discussion #500 and #414 but I can't compile with OpenCV 3.4.0, I get the errors:
1>------ Build started: Project: darknet, Configuration: Debug x64 ------ 1>http_stream.cpp 1>c:\users\winbase\code\opencv\build\include\opencv2\flann\matrix.h(70): warning C4003: not enough actual parameters for macro 'free' 1>c:\users\winbase\code\opencv\build\include\opencv2\flann\matrix.h(70): error C2059: syntax error: ',' 1>c:\users\winbase\code\opencv\build\include\opencv2\flann\matrix.h(87): note: see reference to class template instantiation 'cvflann::Matrix<T>' being compiled 1>c:\users\winbase\code\opencv\build\include\opencv2\core\cuda.hpp(106): error C2059: syntax error: 'constant'
from matrix.h
/** * Convenience function for deallocating the storage data. */ CV_DEPRECATED void free() { fprintf(stderr, "The cvflann::Matrix<T>::free() method is deprecated " "and it does not do any memory deallocation any more. You are" "responsible for deallocating the matrix memory (by doing" "'delete[] matrix.data' for example)"); }
Is there any guidance on which version of OpenCV 3 is compatible with Darknet? I can compile with 2.4.13 but in running
darknet detector train
it always crashes ontry to allocate workspace
. Without OpenCV I can train without problems. I'm not sure how to get the OpenCV compatibility. I'm using Windows 10, Visual Studio 2017, CUDA 9.2, and the following flags:OPENCV;CUDNN;_CRTDBG_MAP_ALLOC;_MBCS;_TIMESPEC_DEFINED;_CRT_SECURE_NO_WARNINGS;_CRT_RAND_S;GPU;WIN32;NDEBUG;_CONSOLE;_LIB;%(PreprocessorDefinitions)
The text was updated successfully, but these errors were encountered: