CoreML Sample App

An starter app that takes a CoreML (.mlmodel) file and can display live classification results from camera input.

Instructions

1. Import model & point to model class

• Export the model from Co-ML and download the .mlmodel file to your computer
• Drag the .mlmodel file into the starter app (under the 1-Model folder)
• In PredictionStatus, set modelObject to the name or your model. For example, a model named Produce.mlmodel would be var modelObject = Produce()

2. Edit Data (JSON & Swift)

• Decide what type of information you want to appear in the app for each label. See 2-Data > mydata.json as an example.
• Edit 2-Data > mydata.json with the labels in your classifier and any associated data you want to appear (e.g., replace “water” or “emoji” with your custom content)
• in 2-Data > Classification.swift, edit based on the format of the data in your JSON file (change names of keys, provide type for value, and provide default values)

2. Edit UI

• Classification UI: Edit 3-UI > PredictionResultView.swift based on the content you want to appear when the classifier identifies an object. You can view a preview of the UI if you open up Canvas in XCode (it will show the default values you added in Classification.swift)
• Launch screen: 3-UI > LaunchScreenView.swift to customize the launch screen contents

Extra Credit: Add an About page

• Edit 3-UI > AboutView.swift to share details about who created the app