05.00 Introduction

Overview

Using pre-trained machine learning models is a great way to get started with machine learning and to create interactive web applications. Although pre-trained models let you get your project up and running quickly, they are limited by what the model was trained to do.

In this module, we will explore the basics of training machine learning models to do specific tasks. We won’t train the models from scratch as that would take a lot of time and require a lot of data. Instead we will use Teachable Machine and ml5.js to retrain models with new data that we provide. This is called transfer learning and it makes it possible to add new abilities to a machine learning model with less input data and training time.

Learning Objectives

• Use Teachable Machine to create a custom image classification machine learning model
• Deploy the custom model into an interactive web app with p5.js
• Use ml5.js to train machine learning models
• Use ml5.js to deploy trained machine learning models

Key Terms

• pre-trained model
• transfer learning
• epochs
• learning loss
• confidence score
• json