Machine Learning & Scientific Python

PyTorch

Updated

Visit website
01 / Overview

What is PyTorch?

PyTorch builds and runs machine-learning models through an open-source tensor and deep-learning framework.

Who is PyTorch for?

Teams evaluating PyTorch for machine learning & scientific python workflows, particularly tensor computation and neural network training. Check the edition and intended deployment before selecting a package.

02 / Capabilities

PyTorch features

The core capabilities to consider for your workflow.

  • Tensor computation
  • Neural network training
  • Model deployment tools
03 / Plans & costs

PyTorch pricing

Plan availability

Open-source software

Review the software licence and distribution, then scope infrastructure, maintenance and support. Commercial hosting, proprietary extensions and managed services can be priced separately.

Calculate the cost of your quote →

Free access & evaluation

An open-source software edition is available. Check the exact licence and distribution; hosting, operation, support and separately licensed extensions may still cost money.

04 / Buying checklist

Before you choose PyTorch

  • Is the task numerical analysis, model training, experiment tracking, model serving or ongoing monitoring?

  • Which languages, hardware, model formats, deployment targets and operating costs must be supported?

05 / Related software

Explore related software

06 / Questions answered

PyTorch FAQs

Does PyTorch have a free plan?

An open-source software edition is available. Check the exact licence and distribution; hosting, operation, support and separately licensed extensions may still cost money.

Who is PyTorch for?

Teams evaluating PyTorch for machine learning & scientific python workflows, particularly tensor computation and neural network training. Check the edition and intended deployment before selecting a package.

What should I compare before buying?

Check the tier that contains your essential features, confirm plan limits and compare the total quote with your own expected usage.