iOS Machine Learning with Core ML, Swift 6, and SwiftUI

In this video, we’ll take a quick look at the project you’ll build throughout the module.

Next, we’ll prepare the image dataset for Create ML. You’ll learn about the expected folder structure, file naming, and format requirements so the data can be used to train our dog-breed classifier model.

We’ll use Create ML to train our custom dog-breed image classifier. You’ll see how to import the dataset, configure training options, monitor progress, and evaluate the model’s performance — all without writing a single line of code.

I'll walk you through the creation of a new SwiftUI app from scratch. You’ll learn how to set up the project, add the trained Create ML model, and explore the automatically generated model class.

We’ll proceed by integrating the trained dog-breed classifier into our SwiftUI app. You’ll see how to create a custom ImageClassifierViewModel class that initializes the Vision Core ML model, sets up a classification request, and prepares everything for running image predictions.

Now that our model is connected, let’s have it return meaningful results. We’ll complete the ImageClassifierViewModel by processing the model’s predictions and formatting them for display in our SwiftUI app.

We’ll build the SwiftUI user interface for the dog-breed classifier app, creating a clean layout that lets users pick an image, run the classification, and view the predicted breed along with its confidence score.

We’ll finish by adding an image picker that lets users choose photos from their library. Once everything’s in place, you can run the app on a simulator or real device and try it out with real dog photos.

A sneak peak at what we’re going to build in this section.

Finally, let's start coding! In this lecture, you're going to create a playground project that recognizes the dominant language of a text.

In this lecture, you're going to implement a string tokenizer that works with text provided in any language.

In this video, you'll explore further features of the NSTagger class. You're going to build a demo that identifies the nouns, verbs, adjectives etc. in a string.

In this lecture, you’re going to implement a demo that can identify names, places and organizations in natural text.

Here's a sneak peak at the synthetic vision demos you'll build in this section.

In this lecture, I'll show you the details of the iOS app that will serve as a starting point for all the demos in this section.

Before we start the actual image analysis implementation, let’s take a closer look at how the Vision framework works.

You're going to implement the Vision request handler in this lecture.

Follow along and implement the image analysis request. You'll use this request to detect rectangular areas in still images.

Vision returns the observations in the Quartz 2D coordinate system. We need to perform a series of transformation before we can visualize the results.

The next step is to visualize the detected observations.

In this demo, you’re going to update the rectangle detector project so that it finds and demarcates regions of text, faces and barcodes in still images.

A sneak peak at the Amazon review sentiment analyzer app you're going to build in this section.

Let's take a look at the training data. This time we'll use the JSON format.

In this lesson, you're going to create a playground and start implementing the logic required to train the sentiment analysis model.

The training process begins by creating an MLTextClassifier instance.

Next, you write the code that exports the trained classifier to a Core ML model file.

In this lecture, we start implementing the Review Sentiment Analyzer iOS app. First, we put together the app’s user interface.

We finish the Review Sentiment Analyzer iOS app by integrating the trained model and implementing the missing functionality.

Finally, let's test the finished Review Sentiment Analyzer app with real product reviews.

Congrats, you’ve reached the end of this course! Let me share some closing thoughts with you.
Thanks for watching!

Here's the companion eBook as a special gift to you (187 pages print length, sells for $28.80 on Amazon).

Some links that will get you discounts on my other courses.