HealthyLuncher was created as a demo project for a CreateML blogpost. It uses:
- CoreML
- NaturalLanguage
- Vision
frameworks in order to demonstrate Machine Learning models for image, text & data table classifications.
HealthyLuncher enables us to predict based on the photo of the lunch whether it is healthy or not. Currently user can chose pictures only from the Photo Library for easier feature testing on the simulator.
User can also classify whether his lunch is healthy or not by typing his lunch in the text screen.
The third data table screen demonstrates the Tabular Data classification.
- Xcode 10.1
- Clone the repository
- Open .xcodeproj file and build the project
You can use HealthyLuncher-Demo for testing your Machine Lerning models. Two things to keep in mind:
- Add new models and change their names if neeeded in ImageClassificationService, TextViewController or DataTableViewController
- Update class labels if needed in Prediction.swift