ubuntu18.04 cuda10.1 pt1.5.1 cudnn8.0.5

1.cuda 10.1(显卡驱动：450，使用的是ubuntu18.04中自带 附加程序，设定的)
https://developer.nvidia.com/cuda-10.2-download-archive?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1804&target_type=runfilelocal

nano ~/.bashrc

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:/usr/local/cuda-10.1/lib64
export PATH=$PATH:/usr/local/cuda-10.1/bin
export CUDA_HOME=$CUDA_HOME:/usr/local/cuda-10.1

source ~/.bashrc

nvcc --version

2.cudnn 8.0.5
https://developer.nvidia.com/rdp/cudnn-archive

2-1根据实际软件包名称，执行下列命令

sudo dpkg -i libcudnn8_8.1.0.77-1+cuda11.2_amd64.deb
sudo dpkg -i libcudnn8-dev_8.1.0.77-1+cuda11.2_amd64.deb
sudo dpkg -i libcudnn8-samples_8.1.0.77-1+cuda11.2_amd64.deb

2-2 确认

cd /usr/src/cudnn_samples_v8/mnistCUDNN/
make clean && make
./mnistCUDNN

若出现下面内容，代表ok

Result of classification: 1 3 5

Test passed!

3.pytorch
https://download.pytorch.org/whl/torch_stable.html
对应cuda 10.1
pytorch1.5.1 - torchvision 0.6.1