1, a new environment, to avoid conflict with the basic environment
conda create -n pytorch python=3.6
2, switch to the environment pytorch
conda activate pytorch
3, the mounting frame substantially
conda install pandas
......
4, installation pytorch
1) Enter the official website https://pytorch.org/get-started/locally/
2) according to their own systems and cuda version, choose the corresponding option
3) Copy the code Run this Command to pytorch in the terminal environment, the installation starts.
4) cuda version View:
a）.windows：
Desktop Right -> NVDIA Control Panel -> Help -> System Information
Components -> NVCUDA.DLL View CUDA version
b）linux：
Open the terminal input: nvcc -V
or:
CUDA:
cat /usr/local/cuda/version.txt
cudnn:
cat /usr/local/cuda/include/cudnn.h | grep CUDNN_MAJOR -A 2