Skip to content

Latest commit

 

History

History
 
 

tfjs-node

README.md

TensorFlow backend for TensorFlow.js via Node.js

This repository provides native TensorFlow execution in backend JavaScript applications under the Node.js runtime, accelerated by the TensorFlow C binary under the hood. It provides the same API as TensorFlow.js.

This package will work on Linux, Windows, and Mac platforms where TensorFlow is supported.

Installing

TensorFlow.js for Node currently supports the following platforms:

For GPU support, [email protected] or later requires the following NVIDIA® software installed on your system:

Name Version
NVIDIA® GPU drivers >450.x
CUDA® Toolkit 11.2
cuDNN SDK 8.1.0

Other Linux variants might also work but this project matches core TensorFlow installation requirements.

Installing CPU TensorFlow.js for Node:

npm install @tensorflow/tfjs-node
(or)
yarn add @tensorflow/tfjs-node

Installing Linux/Windows GPU TensorFlow.js for Node:

npm install @tensorflow/tfjs-node-gpu
(or)
yarn add @tensorflow/tfjs-node-gpu

Windows / Mac OS X Requires Python 2.7

Windows & OSX build support for node-gyp requires Python 2.7. Be sure to have this version before installing @tensorflow/tfjs-node or @tensorflow/tfjs-node-gpu. Machines with Python 3.x will not install the bindings properly.

For more troubleshooting on Windows, check out WINDOWS_TROUBLESHOOTING.md.

Mac OS X Requires Xcode

If you do not have Xcode setup on your machine, please run the following commands:

$ xcode-select --install

For Mac OS Catalina please follow this guide to install node-gyp.

After that operation completes, re-run yarn add or npm install for the @tensorflow/tfjs-node package.

You only need to include @tensorflow/tfjs-node or @tensorflow/tfjs-node-gpu in the package.json file, since those packages ship with @tensorflow/tfjs already.

Rebuild the package on Raspberry Pi

To use this package on Raspberry Pi, you need to rebuild the node native addon with the following command after you installed the package:

$ npm rebuild @tensorflow/tfjs-node --build-from-source

Custom binaries URI

If you happen to be using a mirror for the libtensorflow binaries (default is [https://storage.googleapis.com/]), you have 3 options (in order of priority):

  1. Set the environment variable TFJS_NODE_CDN_STORAGE. This has the same behavior as CDN_STORAGE, but introduced to prevent collisions with other npm packages that might use CDN_STORAGE.
TFJS_NODE_CDN_STORAGE="https://yourmirrorofchoice.com/" npm install <package>
(or)
TFJS_NODE_CDN_STORAGE="https://yourmirrorofchoice.com/" yarn install <package>
  1. Add the variable TFJS_NODE_CDN_STORAGE to your .npmrc file.
TFJS_NODE_CDN_STORAGE=https://yourmirrorofchoice.com/
  1. Set the environment variable CDN_STORAGE. This option is deprecated in favor of the TFJS_NODE_ prefix version above and will be removed in a future release.
CDN_STORAGE="https://yourmirrorofchoice.com/" npm install <package>
(or)
CDN_STORAGE="https://yourmirrorofchoice.com/" yarn install <package>

If your "mirror" uses a custom URI path that doesn't match the default, you have 2 options (in order of priority):

  1. Set the environment variable TFJS_NODE_BASE_URI
TFJS_NODE_BASE_URI="https://yourhost.com/your/path/libtensorflow-" npm install <package>
(or)
TFJS_NODE_BASE_URI="https://yourhost.com/your/path/libtensorflow-" yarn install <package>
  1. Add the variable TFJS_NODE_BASE_URI to your .npmrc file
TFJS_NODE_BASE_URI=https://yourhost.com/your/path/libtensorflow-

Using the binding

Before executing any TensorFlow.js code, import the node package:

// Load the binding
const tf = require('@tensorflow/tfjs-node');

// Or if running with GPU:
const tf = require('@tensorflow/tfjs-node-gpu');

Note: you do not need to add the @tensorflow/tfjs package to your dependencies or import it directly.

Development

# Download and install JS dependencies, including libtensorflow 1.8.
yarn

# Run TFJS tests against Node.js backend:
yarn test
# Switch to GPU for local development:
yarn enable-gpu

MNIST demo for Node.js

See the tfjs-examples repository for training the MNIST dataset using the Node.js bindings.

Optional: Build optimal TensorFlow from source

To get the most optimal TensorFlow build that can take advantage of your specific hardware (AVX512, MKL-DNN), you can build the libtensorflow library from source:

./configure
bazel build --config=opt --config=monolithic //tensorflow/tools/lib_package:libtensorflow

The build might take a while and will produce a bazel-bin/tensorflow/tools/lib_package/libtensorflow.tar.gz file, which should be unpacked and replace the files in deps folder of tfjs-node repo:

cp bazel-bin/tensorflow/tools/lib_package/libtensorflow.tar.gz ~/myproject/node_modules/@tensorflow/tfjs-node/deps
cd path-to-my-project/node_modules/@tensorflow/tfjs-node/deps
tar -xf libtensorflow.tar.gz

If you want to publish an addon library with your own libtensorflow binary, you can host the custom libtensorflow binary and optional pre-compiled node addon module on the cloud service you choose, and add a custom-binary.json file in scripts folder with the following information:

{
  "tf-lib": "url-to-download-customized-binary",
  "addon": {
    "host": "host-of-pre-compiled-addon",
    "remote_path": "remote-path-of-pre-compiled-addon",
    "package_name": "file-name-of-pre-compile-addon"
  }
}

The installation scripts will automatically catch this file and use the custom libtensorflow binary and addon. If addon is not provided, the installation script will compile addon from source.