Maskrcnn-benchmark: error: command '/usr/local/cuda-10.1/bin/nvcc' failed with exit status 1

[ke511081177]

❓ Questions and Help

I in the final step at setup build and follow the readme
but it show the error like below
Have anyone have same problem before~
please help me..

CUDA:10.1
Pytorch:1.4.0
torchvision : 0.5.0
g++ : 5.4
gcc : 5.4

(maskrcnn_benchmark) hank@fdc728d9d94d:~/maskrcnn-benchmark$ CUDAHOSTCXX=/usr/bin/gcc-5 python setup.py build develop
running build
running build_py
creating build
creating build/lib.linux-x86_64-3.8
creating build/lib.linux-x86_64-3.8/maskrcnn_benchmark
copying maskrcnn_benchmark/__init__.py -> build/lib.linux-x86_64-3.8/maskrcnn_benchmark
creating build/lib.linux-x86_64-3.8/maskrcnn_benchmark/utils

......
In file included from /home/hank/maskrcnn-benchmark/maskrcnn_benchmark/csrc/cuda/deform_pool_cuda.cu:9:0:
/home/hank/anaconda3/envs/maskrcnn_benchmark/lib/python3.8/site-packages/torch/include/ATen/cuda/CUDAContext.h:7:23: fatal error: cublas_v2.h: No such file or directory

compilation terminated.

[tuffr5]

1. locate cublas_v2.h
2. cp the file .h to the path to your cuda.
   Rerun

[EthanZhangYC]

I'm using from conda, and adding export PATH=/usr/local/cuda-10.0/bin:$PATH in ~/.bashrc works for me.