hi, I’m currently trying to custom build pytorch. But cuda doesn’t seem to be provided statically. How can I include cuda and cudnn statically?
I have already turned on the following options
USE_CUDA=ON
USE_STATIC_CUDNN=ON
USE_CUDNN=ON
CAFFE2_STATIC_LINK_CUDA
my option
I want the following files to be included in my whl file (only for TORCH_CUDA_ARCH_LIST=8.6)
but my