TensorFlow 2.0 Beta is available Learn more

Training and Prediction in Node.js

In this tutorial you will train a model to make predictions of baseball pitch types from pitch sensor data (from MLBAM). The training will be done server-side in a Node.js application.

This exercise will demonstrate steps to setup the tfjs-node npm package in your server application, build a model, and train it with labeled pitch sensor data. It will also show how to communicate training status to a client, and use the trained model for prediction in a client/server architecture.

Open in CodeLab