A JavaScript implementation of the Web Neural Network API.
The implementation of this webnn-polyfill is based on TensorFlow.js that supports following 4 backends.
- TensorFlow.js CPU Backend, pure-JS backend for Node.js and the browser.
- TensorFlow.js WebGL Backend, WebGL backend for the browser.
- TensorFlow.js WebGPU Backend, WebGPU backend for the browser.
- TensorFlow.js WASM Backend, WebAssembly backend for the browser.
If not set, tests under the webnn-polyfill use CPU as default backend, as which has higher numerical precision than other backends. Tests may fail under WASM backend as some ops have not been implemented/supported in WASM backend.
- For node test, we only support CPU backend.
- For browser test, you can set backend by passing a URL parameter:
backend
, it acceptscpu
,webgl
,webgpu
andwasm
. e.g.?backend=webgl
.
If not set, the built webnn-polyfill.js
uses WebGL as default backend, you can set backend by referring to following code snippet:
const backend = 'cpu';
const tf = navigator.ml.createContext().tf;
await tf.setBackend(backend);
await tf.ready();
** Note: Regarding to browser test using webgpu
backend, please use Chrome Canary browser for Windows platform or Chrome Dev browser for Linux platform with flag "--enable-unsafe-webgpu" enabled.
> git clone --recurse-submodules https://github.com/webmachinelearning/webnn-polyfill
> cd webnn-polyfill & npm install
> npm run build
> npm run build-production
> npm test
> npm start
Open the web browser and navigate to http://localhost:8080/test
> npm run test-cts
> npm start
Open the web browser and navigate to http://localhost:8080/test/cts.html
> npm run build-docs
> npm run lint
> npm run format
> npm run dev
> npm run watch
This project is following Apache License Version 2.0.