In the last part of the book, we covered several on-going projects in the community. These include several new backend implementations that provide us with a chance to use more cutting-edge hardware acceleration technology with TensorFlow.js. The fact that we do not modify the model code encourages us to try and compare several backend environments to find the best one for our application.
Additionally, we introduced a library to run the TensorFlow.js application in native environments (such as mobile- and desktop-based) to expose the application to more users on various kinds of platform. tfjs-react-native enables us to run it with React Native and TensorFlow.js can be run on Electron when we use the tfjs-node backend without any modifications. Try to port your application onto various kinds of platform. This will enable your application to move forward beyond the web...
