Usage

This package adds a WebAssembly backend to TensorFlow.js. It currently supports the following models from our models repo:

• BlazeFace
• BodyPix
• CocoSSD
• Face landmarks detection
• HandPose
• KNN classifier
• MobileNet
• PoseDetection
• Q&A
• Universal sentence encoder
• AutoML Image classification
• AutoML Object detection

Importing the backend

Via NPM

// Import @tensorflow/tfjs or @tensorflow/tfjs-core
import * as tf from '@tensorflow/tfjs';
// Adds the WASM backend to the global backend registry.
import '@tensorflow/tfjs-backend-wasm';
// Set the backend to WASM and wait for the module to be ready.
tf.setBackend('wasm').then(() => main());

Via a script tag

<!-- Import @tensorflow/tfjs or @tensorflow/tfjs-core -->
<script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs/dist/tf.min.js"> </script>

<!-- Adds the WASM backend to the global backend registry -->
<script src="https://cdn.jsdelivr.net/npm/@tensorflow/tfjs-backend-wasm/dist/tf-backend-wasm.js"></script>
<script>
tf.setBackend('wasm').then(() => main());
</script>

Setting up cross-origin isolation

Starting from Chrome 92 (to be released around July 2021), cross-origin isolation needs to be set up in your site in order to take advantage of the multi-threading support in WASM backend. Without this, the backend will fallback to the WASM binary with SIMD-only support (or the vanilla version if SIMD is not enabled). Without multi-threading support, certain models might not achieve the best performance.

Here are the high-level steps to set up the cross-origin isolation. You can learn more about this topic here.

1. Send the following two HTTP headers when your main document (e.g.index.html) that uses the WASM backend is served. You may need to configure or ask your web host provider to enable these headers.

   • Cross-Origin-Opener-Policy: same-origin
   • Cross-Origin-Embedder-Policy: require-corp

2. If you are loading the WASM backend from jsdelivr through the script tag, you are good to go. No more steps are needed.

   If you are loading the WASM backend from your own or other third-party servers, you need to make sure the script is served with either CORS or CORP header.

   • CORS header: Access-Control-Allow-Origin: *. In addition, you will also need to add the "crossorigin" attribute to your script tags.

   • CORP header:

     • If the resource is loaded from the same origin as your main site (e.g. main site: mysite.com/, script: mysite.com/script.js), set:

       Cross-Origin-Resource-Policy: same-origin

     • If the resource is loaded from the same site but cross origin (e.g. main site: mysite.com/, script: static.mysite.com:8080/script.js), set:

       Cross-Origin-Resource-Policy: same-site

     • If the resource is loaded from the cross origin(s) (e.g. main site: mysite.com/, script: mystatic.com/script.js), set:

       Cross-Origin-Resource-Policy: cross-origin

If the steps above are correctly done, you can check the Network tab from the console and make sure the tfjs-backend-wasm-threaded-simd.wasm WASM binary is loaded.

Threads count

By default, the backend will use the number of logical CPU cores as the threads count when creating the threadpool used by XNNPACK. You can use the setThreadsCount API to manually set it (must be called before calling tf.setBackend('wasm')). getThreadsCount API can be used to get the actual number of threads being used (must be called after the WASM backend is initialized).

Via NPM

import * as tf from '@tensorflow/tfjs';
import {getThreadsCount, setThreadsCount} from '@tensorflow/tfjs-backend-wasm';

setThreadsCount(2);
tf.setBackend('wasm').then(() => {
  console.log(getThreadsCount());
});

Via script tag

tf.wasm.setThreadsCount(2);
tf.setBackend('wasm').then(() => {
  console.log(tf.wasm.getThreadsCount());
});

Running MobileNet

async function main() {
  let img = tf.browser.fromPixels(document.getElementById('img'))
      .resizeBilinear([224, 224])
      .expandDims(0)
      .toFloat();

  let model = await tf.loadGraphModel(
    'https://tfhub.dev/google/imagenet/mobilenet_v2_100_224/classification/2',
    {fromTFHub: true});
  const y = model.predict(img);

  y.print();
}
main();

Our WASM backend builds on top of the XNNPACK library which provides high-efficiency floating-point neural network inference operators.

Using bundlers

The shipped library on NPM consists of 2 files:

• the main js file (bundled js for browsers)
• the WebAssembly binary in dist/tfjs-backend-wasm.wasm

There is a proposal to add WASM support for ES6 modules. In the meantime, we have to manually read the wasm file. When the WASM backend is initialized, we make a fetch/readFile for tfjs-backend-wasm.wasm relative from the main js file. This means that bundlers such as Parcel and WebPack need to be able to serve the .wasm file in production. See starter/parcel and starter/webpack for how to setup your favorite bundler.

If you are serving the .wasm files from a different directory, call setWasmPaths with the location of that directory before you initialize the backend:

import {setWasmPaths} from '@tensorflow/tfjs-backend-wasm';
// setWasmPaths accepts a `prefixOrFileMap` argument which can be either a
// string or an object. If passing in a string, this indicates the path to
// the directory where your WASM binaries are located.
setWasmPaths('www.yourdomain.com/');
tf.setBackend('wasm').then(() => {...});

If the WASM backend is imported through <script> tag, setWasmPaths needs to be called on the tf.wasm object:

tf.wasm.setWasmPaths('www.yourdomain.com/');

Note that if you call setWasmPaths with a string, it will be used to load each binary (SIMD-enabled, threading-enabled, etc.) Alternatively you can specify overrides for individual WASM binaries via a file map object. This is also helpful in case your binaries have been renamed.

For example:

import {setWasmPaths} from '@tensorflow/tfjs-backend-wasm';
setWasmPaths({
  'tfjs-backend-wasm.wasm': 'www.yourdomain.com/renamed.wasm',
  'tfjs-backend-wasm-simd.wasm': 'www.yourdomain.com/renamed-simd.wasm',
  'tfjs-backend-wasm-threaded-simd.wasm': 'www.yourdomain.com/renamed-threaded-simd.wasm'
  });
tf.setBackend('wasm').then(() => {...});

If you are using a platform that does not support fetch directly, please set the optional usePlatformFetch argument to true:

import {setWasmPath} from '@tensorflow/tfjs-backend-wasm';
const usePlatformFetch = true;
setWasmPaths(yourCustomPathPrefix, usePlatformFetch);
tf.setBackend('wasm').then(() => {...});

JS Minification

If your bundler is capable of minifying JS code, please turn off the option that transforms typeof foo == "undefined" into foo === void 0. For example, in terser, the option is called "typeofs" (located under the Compress options section). Without this feature turned off, the minified code will throw "_scriptDir is not defined" error from web workers when running in browsers with SIMD+multi-threading support.

Use with Angular

If you see the Cannot find name 'EmscriptenModule' error when building your Angular app, make sure to add "@types/emscripten" to the compilerOptions.types field in your tsconfig.app.json (or tsconfig.json):

{
  ...
  "compilerOptions": {
    "types": [
      "@types/emscripten"
    ]
  },
  ...
}

By default, the generated Angular app sets this field to an empty array which will prevent the Angular compiler from automatically adding "global types" (such as EmscriptenModule) defined in d.ts files to your app.

Benchmarks

The benchmarks below show inference times (ms) for two different edge-friendly models: MobileNet V2 (a medium-sized model) and Face Detector (a lite model). All the benchmarks were run in Chrome 79.0 using this benchmark page across our three backends: Plain JS (CPU), WebGL and WASM. Inference times are averaged across 200 runs.

MobileNet V2

MobileNet is a medium-sized model with 3.48M params and ~300M multiply-adds. For this model, the WASM backend is between ~3X-11.5X faster than the plain JS backend, and ~5.3-7.7X slower than the WebGL backend.

| MobileNet inference (ms) | WASM | WebGL | Plain JS | WASM + SIMD | WASM + SIMD + threads |--------------------------|-------|-------|----------|-------------|---------------------- | iPhone X | 147.1 | 20.3 | 941.3 | N/A | N/A | | iPhone XS | 140 | 18.1 | 426.4 | N/A | N/A | | Pixel 4 | 182 | 76.4 | 1628 | 82 | N/A | | ThinkPad X1 Gen6 w/Linux | 122.7 | 44.8 | 1489.4 | 34.6 | 12.4 | | Desktop Windows | 123.1 | 41.6 | 1117 | 37.2 | N/A | | Macbook Pro 15 2019 | 98.4 | 19.6 | 893.5 | 30.2 | 10.3 | | Node v.14 on Macbook Pro | 290 | N/A | 1404.3 | 64.2 | N/A |

Face Detector

Face detector is a lite model with 0.1M params and ~20M multiply-adds. For this model, the WASM backend is between ~8.2-19.8X faster than the plain JS backend and comparable to the WebGL backend (up to ~1.7X faster, or 2X slower, depending on the device).

| Face Detector inference (ms) | WASM | WebGL | Plain JS | WASM + SIMD | WASM + SIMD + threads |------------------------------|------|-------|----------|-------------|---------------------- | iPhone X | 22.4 | 13.5 | 318 | N/A | N/A | | iPhone XS | 21.4 | 10.5 | 176.9 | N/A | N/A | | Pixel 4 | 28 | 28 | 368 | 15.9 | N/A | | Desktop Linux | 12.6 | 12.7 | 249.5 | 8.0 | 6.2 | | Desktop Windows | 16.2 | 7.1 | 270.9 | 7.5 | N/A | | Macbook Pro 15 2019 | 13.6 | 22.7 | 209.1 | 7.9 | 4.0 |

FAQ

When should I use the WASM backend?

You should always try to use the WASM backend over the plain JS backend since it is strictly faster on all devices, across all model sizes. Compared to the WebGL backend, the WASM backend has better numerical stability, and wider device support. Performance-wise, our benchmarks show that:

• For medium-sized models (~100-500M multiply-adds), the WASM backend is several times slower than the WebGL backend.
• For lite models (~20-60M multiply-adds), the WASM backend has comparable performance to the WebGL backend (see the Face Detector model above).

We are committed to supporting the WASM backend and will continue to improve performance. We plan to follow the WebAssembly standard closely and benefit from its upcoming features such as multi-threading.

How many ops have you implemented?

See register_all_kernels.ts for an up-to-date list of supported ops. We love contributions. See the contributing document for more info.

Do you support training?

Maybe. There are still a decent number of ops that we are missing in WASM that are needed for gradient computation. At this point we are focused on making inference as fast as possible.

Do you work in node?

Yes. If you run into issues, please let us know.

Do you support SIMD and multi-threading?

Yes. We take advantage of SIMD and multi-threading wherever they are supported by testing the capabilities of your runtime and loading the appropriate WASM binary. If you intend to serve the WASM binaries from a custom location (via setWasmPaths), please note that the SIMD-enabled and threading-enabled binaries are separate from the regular binary.

How do I give feedback?

We'd love your feedback as we develop this backend! Please file an issue here.

Development

Emscripten installation

The Emscripten installation necessary to build the WASM backend is managed automatically by the Bazel Emscripten Toolchain.

Building

yarn build

Testing

yarn test

Deployment

./scripts/build-npm.sh
npm publish