PyTorch configuration steps on CentOS

Deploying PyTorch on CentOS system, this article provides detailed steps.

1. Anaconda installation

1. Go to Anaconda official website to download the Anaconda installation package for CentOS.
2. Run the installer and follow the prompts to complete the installation. It is recommended to select the "Just Me" installation method to facilitate subsequent environment variable configuration.

2. Virtual environment creation

1. Open the terminal, create a virtual environment named pytorch_env using the following command and specify the Python version (such as Python 3.10):

 conda create -n pytorch_env python=3.10

2. Activate the virtual environment:

 conda activate pytorch_env

3. Mirror source configuration (optional)

In order to speed up downloads, it is recommended to configure a domestic mirror source. Edit the .condarc file (usually located in the user's home directory) and add the following:

 <code>channels: - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/pytorch/ - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/msys2/ - https://mirrors.tuna.tsinghua.edu.cn/anaconda/pkgs/free/ - https://mirrors.tuna.tsinghua.edu.cn/anaconda/cloud/conda-forges/ show_channel_urls: true auto_activate_base: false</code>

4. PyTorch installation

1. CUDA version check: Use the following command to view the CUDA version:

 nvidia-smi

If CUDA is not installed, please install it according to your requirements.

2. PyTorch download and installation:

   • Visit the PyTorch official website or Tsinghua University open source software mirroring site to obtain the required PyTorch version.
   • Select the PyTorch version that matches the CUDA version. For example, CUDA 12.1 corresponds to CUDA 12.1 version of PyTorch.

   Install using pip (taking CUDA 12.1 as an example):

 pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cu121

Or install using conda :

 conda install pytorch torchvision torchaudio cudatoolkit=12.1 -c pytorch

3. Installation verification: After activating the virtual environment, enter the Python interactive environment and execute the following command to verify the installation:

 import torch
print(torch.__version__)
print(torch.cuda.is_available())

torch.cuda.is_available() returns True to indicate that PyTorch has been successfully installed and can use the GPU.

5. Problem investigation

• NumPy conflicts with torch version: If NumPy and torch version are incompatible, try to downgrade NumPy:

 pip install "numpy<version number>"</version>

• nvcc command not found: nvcc command not found may be due to CUDA installation error or environment variables not configured correctly. Please make sure that the CUDA installation path has been added to the system environment variables.

Refer to the official PyTorch documentation or community forum to solve other problems encountered during the installation process. Hope the above steps can help you configure the PyTorch environment smoothly on CentOS system.

The above is the detailed content of PyTorch configuration steps on CentOS. For more information, please follow other related articles on the PHP Chinese website!