_index.yaml

project_path: /_project.yaml
book_path: /_book.yaml
title: Install TensorFlow 2
landing_page:
  custom_css_path: /site-assets/css/style.css
  nav: both
  meta_tags:
  - name: description
    content: >
      Learn how to install TensorFlow on your system. Download a pip package, run in a Docker
      container, or build from source. Enable the GPU on supported cards.
  rows:
  - classname:
      devsite-landing-row-large-headings
      devsite-landing-row-100
    items:
    - heading: "Install TensorFlow 2"
      description: >
        <p>TensorFlow is tested and supported on the following 64-bit systems:</p>
        <table class="columns">
          <tr><td>
            <ul>
              <li>Python 3.5–3.8</li>
              <li>Ubuntu 16.04 or later</li>
              <li>Windows 7 or later (with <a href="https://support.microsoft.com/help/2977003/the-latest-supported-visual-c-downloads">C++ redistributable</a>)</li>
            </ul>
          </td><td>
            <ul>
              <li>macOS 10.12.6 (Sierra) or later (no GPU support)</li>
              <li>Raspbian 9.0 or later</li>
            </ul>
          </td></tr>
        </table>

  - classname: >
      devsite-landing-row-large-headings
      devsite-landing-row-50
    items:
    - heading: Download a package
      description: >
        <p>Install TensorFlow with Python's <em>pip</em> package manager.</p>
        <aside class="note">TensorFlow 2 packages require a <code>pip</code> version >19.0.</aside>
        <p>Official packages available for Ubuntu, Windows, macOS, and the Raspberry Pi.</p>
        <p>See the <a href="./gpu">GPU guide</a> for CUDA®-enabled cards.</p>

      buttons:
      - label: Read the pip install guide
        path: /install/pip
        classname: button button-primary

      code_block: |
        <pre class="prettyprint lang-bsh devsite-disable-click-to-copy">
        # Requires the latest pip
        <code class="devsite-terminal">pip install --upgrade pip</code><br/>
        # Current stable release for CPU and GPU
        <code class="devsite-terminal">pip install tensorflow</code><br/>
        # Or try the preview build (unstable)
        <code class="devsite-terminal">pip install tf-nightly</code><br/>

  - classname:
      devsite-landing-row-100
      devsite-landing-row-large-headings
    items:
    - heading: Run a TensorFlow container
      description: >
        <p>
          The <a href="https://hub.docker.com/r/tensorflow/tensorflow/" class="external">TensorFlow
          Docker images</a> are already configured to run TensorFlow. A
          <a href="https://docs.docker.com/install/" class="external">Docker</a> container runs in a
          virtual environment and is the easiest way to set up <a href="/install/gpu">GPU
          support</a>.
        </p>
        <pre class="prettyprint lang-bsh">
        <code class="devsite-terminal">docker pull tensorflow/tensorflow:latest  # Download latest stable image</code><br/>
        <code class="devsite-terminal">docker run -it -p 8888:8888 tensorflow/tensorflow:latest-jupyter  # Start Jupyter server</code>
        </pre>
      buttons:
      - label: Read the Docker install guide
        path: /install/docker
        classname: button button-primary

  - classname:
      devsite-landing-row-large-headings
      devsite-landing-row-100
    items:
    - heading: Google Colab&#58; An easy way to learn and use TensorFlow
      description: >
        <p>
          No install necessary—run the <a href="/tutorials">TensorFlow tutorials</a> directly in
          the browser with <a href="https://colab.research.google.com/notebooks/welcome.ipynb"
          class="external">Colaboratory</a>, a Google research project created to help disseminate
          machine learning education and research. It's a Jupyter notebook environment that requires
          no setup to use and runs entirely in the cloud.
          <a href="https://medium.com/tensorflow/colab-an-easy-way-to-learn-and-use-tensorflow-d74d1686e309" class="external">Read the blog post</a>.
        </p>

  - background: grey
    heading: Build your first ML app
    description: Create and deploy TensorFlow models on web and mobile.
    items:
    - classname: tfo-landing-row-item-inset-white
      heading: Web developers
      path: https://js.tensorflow.org
      description: >
        TensorFlow.js is a WebGL accelerated, JavaScript library to train and
        deploy ML models in the browser and for Node.js.
    - classname: tfo-landing-row-item-inset-white
      heading: Mobile developers
      path: /lite/
      description: >
        TensorFlow Lite is a lightweight solution for mobile and embedded devices.