TensorFlow.js Example: Training a baseball model in Node.js

This demo demonstrates how to train a server-side model to classify baseball pitch types using Node.js.

It has four parts:

Baseball sensor training and test data.
Two ML models that do classification given the sensor data:
- Model that predicts the type of pitch.
- Model that predicts if there was a strike.
Node.js server that trains a model and serves results over a web socket.
Web application that displays pitch type learning statistics.

Running the Demo

First, prepare the environment:

$ npm install
# or
$ yarn

Next, start the client:

$ npm run start-client
# or
$ yarn start-client

Open the client running at: http://localhost:8080/

In a new shell, start the server:

$ npm run start-server
# or
$ yarn start-server

Two small scripts are provided to test training both of the baseball models without running the client/server demo:

Pitch Type model:

$ node train_pitch_type.js

Strike Zone model:

$ node train_strike_zone.js

Baseball Models

This demo contains two models. The first is a pitch-type model used in the actual client/server architecture. The other model learns how to call balls and strikes like a major-league umpire. It currently does not have any presentation UI but exists for developers to experiment with.

Pitch type model - Classifies 7 different pitch types looking at baseball sensor data (pitch-type-model.ts)
Strike zone model - A model that can learns how to call balls and strikes based on historical umpire calls (strike-zone-model.ts).