Open sangheonEN opened 2 years ago
Sorry, we don't support windows, the bottleneck is compilation. You may need to modify the compiled script, but I'm not familiar with that.
If so, it seems that I can implement it myself with py code by referring to the cpp code.
/alpharotate/libs/utils/iou_cpu.cpp /alpharotate/libs/utils/rbbox_overlaps.cpp /alpharotate/libs/utils/rotate_gpu_nms.hpp /alpharotate/libs/utils/rotate_polygon_nms.cpp
And I don't find the /alpharotate.libs.box_utils folder!
alpharotate/libs/utils
alpharotate/libs/utils
Likewise, I don't see the box_utils above path.
I rename it alpharotate/libs/utils. where do you find alpharotate/libs/box_utils
207 line in iou_rotate.py -> from alpharotate.libs.box_utils.overlaps_cython.overlaps_cython import rbox_overlaps
is rbox_overlaps function name change?
what is rbox_overlaps name now?
What path should I set each of those cudaconfig parameters to in the CUDA Toolkit?
In general, when setting the path to the cuda toolkit environment variable, is that nvcc path the nvcc.exe file in cudatoolkit?
Hi @sangheonEN were you able to rewrite the compilation for windows ? I'm also finding difficulty in compiling it on windows. Help out
What am i solve the problem. Please tell me about the problem. thank you.