pytorch-domains.html

---
layout: default
title: PyTorch Domains
permalink: /pytorch-domains
body-class: blog
background-class: features-background
---

<div class="jumbotron jumbotron-fluid contributor-jumbotron">
  <div class="contributor-jumbo-text container">
    <h1>PyTorch Domains</h1>
    <h2 class="lead">Learn about the various parts of the PyTorch ecosystem</h2>

  </div>
</div>

<div class="main-content-wrapper">
  <div class="main-content">
    <div class="container">
      <div class="row mt-5 justify-content-around">
        <div class="col-sm-5 domain-card">
          <a href="https://pytorch.org/audio/stable/index.html">
            <h4>torchaudio</h4>
            <p>Torchaudio is a library for audio and signal processing with PyTorch. It provides I/O, signal and data processing functions, datasets, model implementations and application components.
            </p>
          </a>
        </div>
        <div class="col-sm-5 domain-card">
          <a href="https://pytorch.org/torchtune/">
            <h4>torchtune</h4>
            <p>A PyTorch-native library designed for fine-tuning large language models (LLMs). torchtune provides supports the full fine-tuning workflow and offers compatibility with popular production inference systems.
            </p>
          </a>
        </div>
        <div class="col-sm-5 domain-card">
          <a href="https://pytorch.org/vision/stable/index.html">
            <h4>torchvision</h4>
            <p>This library is part of the PyTorch project. PyTorch is an open source machine learning framework. The torchvision package consists of popular datasets, model architectures, and common image transformations for computer vision.
            </p>
          </a>
        </div>
        <div class="col-sm-5 domain-card">
          <a href="https://pytorch.org/torcharrow/beta/index.html">
            <h4>torcharrow</h4>
            <p>TorchArrow is a torch.Tensor-like Python DataFrame library for data preprocessing in deep learning. It supports multiple execution runtimes and Arrow as a common format.
            </p>
          </a>
        </div>
        <div class="col-sm-5 domain-card">
          <a href="https://pytorch.org/data/beta/index.html">
            <h4>TorchData</h4>
            <p>A Beta library of common modular data loading primitives for easily constructing flexible and performant data pipelines. And, there are a few features still in prototype stage.
            </p>
          </a>
        </div>
        <div class="col-sm-5 domain-card">
          <a href="https://pytorch.org/torchrec/">
            <h4>TorchRec</h4>
            <p>TorchRec is a PyTorch domain library built to provide common sparsity & parallelism primitives needed for large-scale recommender systems (RecSys). It allows authors to train models with large embedding tables sharded across many GPUs.
            </p>
          </a>
        </div>
        <div class="col-sm-5 domain-card">
          <a href="https://pytorch.org/serve/">
            <h4>TorchServe</h4>
            <p>TorchServe is a performant, flexible and easy to use tool for serving PyTorch models in production.
            </p>
          </a>
        </div>
        <div class="col-sm-5 domain-card">
          <a href="https://pytorch.org/xla/release/2.1/index.html">
            <h4>PyTorch on XLA Devices</h4>
            <p>PyTorch runs on XLA devices, like TPUs, with the torch_xla package. This document describes how to run your models on these devices.
            </p>
          </a>
        </div>
        <div class="col-sm-5 domain-card">
          <a href="https://pytorch.org/rl">
            <h4>torchrl</h4>
            <p>torchrl is an open-source, Python-first Reinforcement Learning library for PyTorch with a focus on high modularity and good runtime performance, providing low and high-level RL abstractions and reusable functionals for cost functions, returns, and data processing.
            </p>
          </a>
        </div>
        <div class="col-sm-5 domain-card">
          <a href="https://pytorch.org/tensordict">
            <h4>tensordict</h4>
            <p>tensordict is a dictionary-like class that simplifies operations on batches of tensors, enhancing code readability, compactness, and modularity by abstracting tailored operations and reducing errors through automatic operation dispatching on the leaves.
            </p>
          </a>
        </div>
        <div class="col-sm-5 domain-card">
          <a href="https://pytorch.org/text/stable/index.html">
            <h4>torchtext</h4>
            <p>This library is part of the PyTorch project. PyTorch is an open source machine learning framework. The torchtext package consists of data processing utilities and popular datasets for natural language.
            </p>
          </a>
        </div>
      </div>
    </div>
  </div>
</div>