Hackathon Workshop 2025: iOS CreateML Audio Classifier with GIPHY API Integration.

Welcome to the CreateML Audio Classifier Workshop! 🎵 This hands-on exercise will guide you through creating an iOS app that classifies audio genres using CreateML and enhances the user experience by integrating the GIPHY API for visual feedback.

---

🛠️ Workshop Overview

In this workshop, participants will:

1. Train an Audio Classifier Model: Use the CreateML framework and the GTZAN dataset to build a machine learning model that classifies audio genres.

2. Integrate the Model into an iOS App: Process live microphone input to classify audio genres in real time.

3. Enhance UX with GIPHY: Dynamically fetch and display GIPHY stickers related to the identified audio genre.

---

📂 Repository Structure

This repository contains two branches:

• **starter**: The starting point for the workshop. Participants will complete key areas of the code marked with //TODO: Complete this.

• **main**: The complete solution for reference.

Clone the branch of your choice:

# Clone the starter branch to follow along
git clone -b starter https://github.com/Adolfo-David-Romero/Workshop-CreateML.git

# Clone the main branch to see the complete solution
git clone -b main https://github.com/Adolfo-David-Romero/Workshop-CreateML.git

---

🧰 Prerequisites

Before you begin, ensure you have the following:

1. Xcode 14+ installed on macOS.

2. Basic knowledge of Swift and SwiftUI.

3. CreateML: Pre-installed as part of Xcode.

4. GIPHY API Key: Sign up at GIPHY Developers to get an API key.

5. GTZAN Dataset: Download the dataset from GTZAN for training the model.

---

🚀 Getting Started

1. Train the Audio Classifier

• Open CreateML in Xcode.

• Use the GTZAN dataset to create an audio classification model that distinguishes genres like Rock, Jazz, Pop, etc.

• Export the trained .mlmodel file and add it to the Resources folder of the Xcode project.

2. Follow the Workshop

• Open the starter branch in Xcode.

• Complete the //TODO: Complete this, or //FIXME: Fix this sections in the code, which cover:

  • Initializing the CoreML model.

  • Processing live audio input into the correct format for predictions.

  • Performing predictions and handling results.

  • Fetching genre-specific GIPHY stickers via the GIPHY API.

• Use the provided PredictionView and HomeView to test the app.

3. Run the App

• Build and run the app on a simulator or physical device.

• Speak or play audio near the microphone and watch the app classify the genre in real-time.

• View related GIPHY stickers displayed dynamically at the bottom of the screen.

---

🌟 Key Features

• Real-Time Audio Classification: Uses CoreML and CreateML for live audio analysis.

• Dynamic Sticker Display: Integrates the GIPHY API to fetch and display stickers based on detected genres.

• Reusable UI Components: Modularized SwiftUI views for easy customization.

---

📦 Dependencies

The project uses the following:

• CreateML: For training the audio classifier.

• CoreML: For integrating the trained model into the app.

• GIPHY API: For fetching genre-specific stickers.

---

📚 Learnings

By the end of this workshop, participants will:

1. Understand how to train a machine learning model with CreateML.

2. Integrate a CoreML model into an iOS app for live predictions.

3. Use a third-party API (GIPHY) to enhance user experience.

4. Follow best practices for modular SwiftUI and MVVM architecture.

---

🤝 Contributing

Feel free to open issues or submit pull requests to improve this repository. Contributions are always welcome!

---

📝 License

This project is licensed under the MIT License.

---

🔗 Resources

• CreateML Documentation

• CoreML Documentation

• GIPHY API Documentation

• GTZAN Dataset