_index.yaml

book_path: /overview/_book.yaml
project_path: /overview/_project.yaml
title: Guide
landing_page:
  custom_css_path: /site-assets/css/style.css
  nav: left
  meta_tags:
  - name: description
    content: >
      Learn basic and advanced concepts of TensorFlow such as eager execution, Keras
      high-level APIs and flexible model building.
  rows:
  - classname:
      devsite-landing-row-100
      devsite-landing-row-large-headings
    items:
    - description: >
        <p>
        TensorFlow 2 focuses on simplicity and ease of use, with updates like
        eager execution, intuitive higher-level APIs, and flexible model
        building on any platform.
        </p>
        <p>
        Many guides are written as Jupyter notebooks and run directly in
        Google&nbsp;Colab—a hosted notebook environment that requires no setup.
        Click the <em>Run in Google&nbsp;Colab</em> button.
        </p>

  - heading: Essential documentation
    classname: tfo-landing-page-heading
    items:
    - classname: tfo-landing-page-card
      description: >
        <a href="/install"><h3 class="no-link">Install TensorFlow</h3></a>
        Install the package or build from source. GPU support for
        CUDA®-enabled cards.
      path: /install
    - classname: tfo-landing-page-card
      description: >
        <a href="/guide/migrate"><h3 class="no-link">Migrate to TensorFlow 2</h3></a>
        Learn how to migrate your TF1.x code to TF2.
      path: /guide/migrate
    - classname: tfo-landing-page-card
      description: >
        <a href="/guide/keras/overview"><h3 class="no-link">Keras</h3></a>
        Keras is a high-level API that's easier for ML beginners, as well as
        researchers.
      path: /guide/keras/overview

  - items:
    - classname: tfo-landing-page-card
      description: >
        <a href="/guide/eager"><h3 class="no-link">TensorFlow basics</h3></a>
        Learn about the fundamental classes and features that make TensorFlow
        work.
      path: /guide/eager
    - classname: tfo-landing-page-card
      description: >
        <a href="/guide/data"><h3 class="no-link">Data input pipelines</h3></a>
        The <code>tf.data</code> API enables you to build complex input pipelines
        from simple, reusable pieces.
      path: /guide/data
    - classname: tfo-landing-page-card
      description: >
        <a href="/guide/effective_tf2"><h3 class="no-link">TensorFlow 2 best practices</h3></a>
        Learn about the best practices for effective development using TensorFlow 2.
      path: /guide/effective_tf2

  - items:
    - classname: tfo-landing-page-card
      description: >
        <a href="/guide/checkpoint"><h3 class="no-link">Save a model</h3></a>
        Save a TensorFlow model using checkpoints or the SavedModel format.
      path: /guide/checkpoint
    - classname: tfo-landing-page-card
      description: >
        <a href="/guide/distributed_training"><h3 class="no-link">Accelerators</h3></a>
        Distribute training across multiple GPUs, multiple machines or TPUs.
      path: /guide/distributed_training
    - classname: tfo-landing-page-card
      description: >
        <a href="/guide/function"><h3 class="no-link">Performance</h3></a>
        Best practices and optimization techniques for optimal TensorFlow performance.
      path: /guide/function

  - classname: devsite-landing-row-100
    items:
    - description: >
        <h2 class="tfo-landing-page-heading no-link">Libraries and extensions</h2>
        Explore <a href="/resources">additional resources</a> to build advanced
        models or methods using TensorFlow, and access domain-specific
        application packages that extend TensorFlow.

  - classname: devsite-landing-row-100
    items:
    - list:
      - description: >
          <a href="/decision_forests"><h3 class="tfo-landing-page-heading no-link">TensorFlow Decision Forests
          </h3></a>
          A library to train, run and interpret decision forest models (e.g., Random Forests,
          Gradient Boosted Trees) in TensorFlow.
        path: /decision_forests
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="/hub"><h3 class="tfo-landing-page-heading no-link">TensorFlow Hub</h3></a>
          A library for the publication, discovery, and consumption of reusable
          parts of machine learning models.
        path: /hub
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="/tfx/guide/serving"><h3 class="tfo-landing-page-heading no-link">Serving</h3></a>
          A TFX serving system for ML models, designed for high-performance in
          production environments.
        path: /tfx/guide/serving
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="/federated"><h3 class="tfo-landing-page-heading no-link">TensorFlow Federated</h3></a>
          A framework for machine learning and other computations on decentralized
          data.
        path: /federated
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="/neural_structured_learning"><h3 class="tfo-landing-page-heading no-link">Neural Structured Learning</h3></a>
          A learning paradigm to train neural networks by leveraging structured
          signals in addition to feature inputs.
        path: /neural_structured_learning
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="/graphics"><h3 class="tfo-landing-page-heading no-link">TensorFlow Graphics</h3></a>
          A library of computer graphics functionalities ranging from cameras, lights, and materials to renderers.
        path: /graphics
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="https://github.com/tensorflow/addons"><h3 class="tfo-landing-page-heading no-link">SIG Addons</h3></a>
          Extra functionality for TensorFlow, maintained by SIG&nbsp;Addons.
        path: https://github.com/tensorflow/addons
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
    - list:
      - description: >
          <a href="/tensorboard"><h3 class="tfo-landing-page-heading no-link">TensorBoard</h3></a>
          A suite of visualization tools to understand, debug, and optimize
          TensorFlow programs.
        path: /tensorboard
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="/datasets"><h3 class="tfo-landing-page-heading no-link">Datasets</h3></a>
          A collection of datasets ready to use with TensorFlow.
        path: /datasets
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="/model_optimization"><h3 class="tfo-landing-page-heading no-link">Model Optimization</h3></a>
          The TensorFlow Model Optimization Toolkit is a suite of tools for
          optimizing ML models for deployment and execution.
        path: /model_optimization
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="/probability"><h3 class="tfo-landing-page-heading no-link">Probability</h3></a>
          TensorFlow Probability is a library for probabilistic reasoning and
          statistical analysis.
        path: /probability
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="/mlir"><h3 class="tfo-landing-page-heading no-link">MLIR</h3></a>
          MLIR unifies the infrastructure for high-performance ML models in
          TensorFlow.
        path: /mlir
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="/xla"><h3 class="tfo-landing-page-heading no-link">XLA</h3></a>
          A domain-specific compiler for linear algebra that accelerates
          TensorFlow models with potentially no source code changes.
        path: /xla
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey
      - description: >
          <a href="https://github.com/tensorflow/io"><h3 class="tfo-landing-page-heading no-link">SIG IO</h3></a>
          Dataset, streaming, and file system extensions, maintained by SIG&nbsp;IO.
        path: https://github.com/tensorflow/io
        icon:
          icon_name: chevron_right
          foreground: theme
          background: grey