PyTorch is a Torch-based Python open source machine learning library for applications such as natural language processing. It is mainly developed by Facebookd's artificial intelligence team. It not only enables powerful GPU acceleration, but also supports dynamic neural networks, which is not supported by many mainstream frameworks such as TensorFlow. This article details the steps to install Pytorch and Pycharm configuration.
After installing Anaconda, use the command line mode
to create an environment
conda create -n pytorch python=3.7
However, it always reported that the https connection could not be connected
. I tried many methods but could not solve it. Finally, I used this method to solve the problem.
The original address https://github.com/conda/conda/issues/8273
is to the effect: conda found the wrong address of openssl . conda looks for the openssl dll file in the Anaconda\DLLs directory , but the actual required dll is in the Anaconda3\library\bin directory. So you only need to copy these two files to Anaconda\DLLs.
Follow the prompts to copy the two dlls to the specified directory.
D:\tools\anaconda\Library\bin —>D:\tools\anaconda\DLLs
Create the environment again conda create -n pytorch python=3.7 -n The following is the name of the environment. To customize the
selection,
use the command, conda env list to view Environment
Then we enter the command to enter the pytorch environment
conda activate pytorch
In order to improve download speed, use domestic download sources
Tsinghua source:
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/main/
conda config --add channels https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/
conda config --set show_channel_urls yes
Dayuan of Science and Technology of China
conda config --add channels https://mirrors.ustc.edu.cn/anaconda/pkgs/main/
conda config --add channels https://mirrors.ustc.edu.cn/anaconda/pkgs/free/
conda config --add channels https://mirrors.ustc.edu.cn/anaconda/cloud/conda-forge/
conda config --add channels https://mirrors.ustc.edu.cn/anaconda/cloud/msys2/
conda config --add channels https://mirrors.ustc.edu.cn/anaconda/cloud/bioconda/
conda config --add channels https://mirrors.ustc.edu.cn/anaconda/cloud/menpo/
conda config --set show_channel_urls yes
After entering each line of command, use the command to check whether the new download source takes effect.
conda config --show-source
Next install Pytorch and find the version that suits you on the official website, https://pytorch.org/get-started/previous-versions/
conda install pytorch==1.11.0 torchvision==0.12.0 torchaudio==0.11.0 cudatoolkit=11.3 -c pytorch
It takes a long time. When this page finally appears, it means the installation is complete.
Exit the pytorch environment.
conda deactivate
View environmental information
conda info -e
Verify that pytorch is installed successfully.
Enter the pytorch environment
and import torch.
torch.cuda.is_available() verifies whether cuda is running normally. I have not installed cuda, so this is False
to uninstall the pytorch environment.
conda remove -n pytorch --all
Configure pycharm
Install ipython and enter the environment
conda install nb_conda
After installation enter the command
jupyter notebook
Jupyter will pop up automatically. You can also enter the following code inside to verify whether the torch is successfully installed and whether the GPU can be used in this environment.
import torch
torch.cuda.is_available()
torch.cuda.is_available() displays False.
Check pytorch1.11 to see how much the driver version needs to be.
If our driver version is smaller than this value, we need to upgrade the driver.
nvidia-smi
There are two ways: one is to upgrade through various software managers, the other is to find the corresponding driver through https://www.nvidia.cm, download it, install it directly, and then come back to see if the driver version is greater than