At that moment on-device AI demons cannot be shared, but according to Meta, that functionality is on its way. Developers have already the ability to start building custom machine learning models to later share with the broader PyTorch community.
In January 2017 PyTorch was publicly launched by Meta. At that time the company was still known as Facebook and the open-source machine learning library turned out to be a firm favourite among the developer data science communities.
The main library’s interface’s design is inspired by Python, but it also includes a C++ interface, as the name suggests.
Previously, TensorFlow was a dominant machine learning but it has been falling behind in usage in recent years. PyTorch’s growth was highlighted in GitHub’s Octoverse in 2018, as an open-source project outpacing that of TensorFlow.
It seems that PyTorch Live intends to accelerate the success of the machine learning library. React Native is used by the tools for building cross-platform visual user interfaces and PyTorch Mobile powers on-device inference.
If you would like to get started with PyTorch Live, you can do so through its command-line interface setup and/or its data processing API.