When i have compile the Deformable Convolutional Networks V2(DCNv2),it comes an error of " fatal error: cuda_runtime_api.h: No such file or directory",and thus i open the path “usr/local/cuda/” and “usr/local/cuda-10.1/” ,and i can’t find the “include” document,so what should i do for the problem?
compile problem
Hi, is cuda installed on your system? Try which nvcc to check.
Your cuda_runtime_api.h should be located similar to this:
CUDA_PATH/bin/nvcc
CUDA_PATH/include/cuda_runtime_api.h
You can provide more info about how you compile the megengine, like compile steps, system version …
CPATH env may help you, ref https://github.com/MegEngine/MegEngine/blob/master/scripts/cmake-build/host_build.sh#L178