How to Check PyTorch Version

Here you will learn how to check PyTorch version in Python or from command line through your Python package manager pip or conda (Anaconda/Miniconda).


You should have installed PyTorch already, which is the assumption of this tutorial. If you have not install PyTorch, search install PyTorch — we have written a bunch of tutorial on this for various versions.

Use Python code to check PyTorch version

If you are in the Python interpreter or want to use programmingly check PyTorch version, use torch.__version__. Note that if you haven’t import PyTorch, you need to use import torch in the beginning of your Python script or before the print statement below.

import torch

check pytorch version in python interpreter

Use pip to check PyTorch package version

If you have used pip to install PyTorch, you can use pip3 show to check the details of PyTorch. Note that pip show may also work.

pip3 show torch

You will see something like below. The second line starting with version will show which version have you installed or updated PyTorch. Here I have installed 1.5.1. The +cu101 means my cuda version is 10.1. This is because I am running Ubuntu 20.04, which comes with CUDA 10.1 by default. We wrote a tutorial before on how to install PyTorch on Ubuntu 20.04.

check pytorch version using pip3 show torch

Here is the full output in text for your reference:

vh@varhowto-com:~$ pip3 show torch
Name: torch
Version: 1.5.1+cu101
Summary: Tensors and Dynamic neural networks in Python with strong GPU acceleration
Author: PyTorch Team
License: BSD-3
Location: /home/vh/.local/lib/python3.8/site-packages
Requires: future, numpy
Required-by: torchvision

Use conda to check PyTorch package version

Similar to pip, if you used Anaconda to install PyTorch. you can use the command conda list to check its detail which also include the version info.

conda list -f pytorch

You you want to check in another environment, e.g., pytorch14 below, use -n like this:

conda list -n pytorch14 -f pytorch

What is PyTorch?

PyTorch is an open-source Deep Learning framework for research, stable and enabling implementation that is scalable and flexible. This enables quick, scalable testing through an autograding component designed for fast and python-like execution. The framework now has graph-based execution with the release of PyTorch 1.0, a hybrid front-end that allows for seamless mode switching, interactive monitoring, and efficient and stable implementation on mobile platforms.

PyTorch has 4 key features according to its official homepage.

  1. PyTorch is production-ready: TorchScript smoothly toggles between eager and graph modes. TorchServe speeds up the production process.
  2. PyTorch support distributed training: The torch.collaborative interface allows for efficient distributed training and performance optimization in research and development.
  3. PyTorch has a robust ecosystem: It has an expansive ecosystem of tools and libraries to support applications such as computer vision and NLP.
  4. PyTorch has native cloud support: It is well recognized for its zero-friction development and fast scaling on key cloud providers.

[summary] 3 Ways to Check PyTorch Version

  1. [Python] Write Python code to check PyTorch version

    You can use torch.__version__ to check the version of PyTorch. If you have not imported PyTorch, use import torch first.check pytorch version in python interpreter

  2. [pip] Use pip3 to check the PyTorch package information

    If you used pip to install PyTorch, run pip3 show torch to show all the information of the installation, which also includes the version of PyTorch.check pytorch version using pip3 show torch

  3. [conda] Use conda list to show the PyTorch package information

    If you used Anaconda or Miniconda to install PyTorch, you can use conda list -f pytorch to check PyTorch package's information, which also includes its version.

    If you want to check PyTorch version for a specific environment such as pytorch14, use conda list -n pytorch14 -f pytorch.


By VarHowto Editor

Welcome to VarHowto!