Creating a Neural Network Based MAUI Mobile App - Part 1 (Google Colab)

Bringing his AI masters degree back into his .NET MAUI work, Stephen Moreton-Howell walks through training a simple neural network in Google Colab and deploying it into a mobile app using ONNX.

What you’ll learn

• The tooling — using Python, Keras, NumPy, and Google Colab to build and train a network
• A worked example — classifying lists of ten numbers as low, medium, or high from generated training data
• Building the model — defining a Sequential Keras model, compiling, training, and testing it
• Exporting for mobile — saving the trained model as .tflite and .onnx files
• Running it in .NET MAUI — installing Microsoft.ML.OnnxRuntime and using an OnnxModelInference to make predictions in-app

Read the full post for the complete Colab and MAUI code, and continue to part 2.