1. Right-click on the desktop and select "NIVIDIA Control Panel". If you don’t have this option, find another way to solve it.
2. In the NIVIDIA control panel, check the CUDA version of our computer. Mine is version 10.1.
3. Enter the Pytorch official website. , Link: https://pytorch.org/
select "Get Started"
4. Check and configure according to your own computer version and other needs. Students who don’t have Anaconda need to install Anaconda separately
. "Run this Command" The column will adjust the command according to your choice, and finally copy this command to cmd to install
6. The installation is successful, as shown in the figure, it will be installed at the same time: future, torch, torchversion
7. The way to test the successful installation:
enter "import torch" in the .py script of pycharm, the installation is successful if there is no error