apple/coremltools

What it solves

Core ML Tools provides a way to take machine learning models trained in popular third-party libraries and convert them into the Core ML format. This allows developers to integrate these models into Apple apps, enabling on-device predictions and fine-tuning without requiring a network connection, which improves privacy and app responsiveness.

How it works

It is a Python package that acts as a converter and optimizer. It supports converting models from frameworks like TensorFlow (1.x and 2.x), PyTorch, scikit-learn, XGBoost, and LibSVM. Once converted, the model is saved in the Core ML format, which can then be integrated into an app via Xcode.

Who it’s for

Developers who want to deploy machine learning models on Apple devices by leveraging the same model format required by the Apple Core ML framework.

Highlights

• Convert models from a wide range of neural network and non-neural network frameworks.
• Read, write, and optimize Core ML models.
• Verify conversion results by making predictions on macOS.
• Enables hardware acceleration via the CPU, GPU, and Neural Engine on Apple devices.