likedan/awesome-coreml-models

This repository is a collection of machine learning models in Apple's Core ML format, built for developers who want to add AI-powered features to their iOS, macOS, tvOS, or watchOS apps. Apple introduced Core ML with iOS 11, and it allows apps to run machine learning models directly on the device without sending data to a server. This collection aims to be the largest single place to find ready-to-use Core ML models. The models in the collection cover a wide range of tasks. Many work with images: some identify what objects appear in a photo, others predict the location where a photo was taken, classify the type of food shown, detect flowers, recognize artistic styles, or estimate a person's age and gender from a portrait. There are models for detecting text in images, predicting the depth of a scene from a single photo, and segmenting a camera frame into labeled regions like road, sky, or person. Beyond images, the collection includes models for natural language tasks. These include sentiment analysis of text, a tool for predicting whether a document is toxic, and models that generate text responses. There is also a section for style transfer models, which repaint a photo to look like a painting, and a section for audio models that can identify the genre of a song or recognize spoken words. Each entry in the collection links to a downloadable model file, a demo app that shows the model working on a real device, and a reference paper or source. Developers can plug the downloaded file directly into an Xcode project. The repository also mentions Netron, a third-party viewer that lets you inspect the structure of a model file. The project is community-maintained and accepts contributions. If someone has converted a new model to Core ML format, they can submit a pull request to add it to the list.