Welcome to the TorchRec documentation!#

TorchRec is a specialized library within the PyTorch ecosystem, tailored for building, scaling, and deploying large-scale recommendation systems, a niche not directly addressed by standard PyTorch. TorchRec offers advanced features such as complex sharding techniques for massive embedding tables, and enhanced distributed training capabilities.

Getting Started#

Topics in this section will help you get started with TorchRec.

TorchRec Overview

A short intro to TorchRec and why you need it.

overview.html

Set up TorchRec

Learn how to install and start using TorchRec in your environment.

setup-torchrec.html

Getting Started with TorchRec Tutorial

Follow our interactive step-by-step tutorial to learn how to use TorchRec in a real-life example.

https://colab.research.google.com/github/pytorch/torchrec/blob/main/Torchrec_Introduction.ipynb

How to Contribute#

We welcome contributions and feedback from the PyTorch community! If you are interested in helping improve the TorchRec project, here is how you can contribute:

Visit Our GitHub Repository: There you can find the source code, issues, and ongoing projects.

Submit Feedback or Issues: If you encounter any bugs or have suggestions for improvements, please submit an issue through the GitHub issue tracker.

Propose changes: Fork the repository and submit pull requests. Whether it’s fixing a bug, adding new features, or improving documentation, your contributions are always welcome! Please make sure to review our CONTRIBUTING.md