john-rocky/CoreML-Models
Converted CoreML Model Zoo.
What it solves
It provides a centralized "model zoo" of pre-converted CoreML models, eliminating the need for developers to manually convert complex AI models from other frameworks (like PyTorch or TensorFlow) to run natively on Apple devices.
How it works
The repository curates a wide variety of state-of-the-art models across different AI domains. Users can browse the categories, download the .mlpackage files via provided Google Drive or GitHub release links, and bundle them directly into their Xcode projects. Some models include links to sample projects to demonstrate implementation.
Who it’s for
iOS, macOS, and iPadOS developers who want to integrate machine learning capabilities into their apps using Apple's CoreML framework without performing the conversion process themselves.
Highlights
- Diverse Modalities: Includes models for image classification, object detection (YOLO series), segmentation (SAM2), super-resolution, and image generation.
- LMM Support: Provides CoreML versions of modern language models like Gemma 4 and Qwen3.5.
- Audio & Speech: Features models for text-to-speech (Kokoro-82M), voice conversion, and audio source separation.
- Ready-to-Use: Offers direct download links and, in many cases, sample implementation projects to accelerate development.