CoreML-Models

Why we included this project

For iOS and macOS developers who run models on-device, this repo saves the most tedious step: converting PyTorch or TensorFlow checkpoints into CoreML format. The models come pre-converted as .mlmodel files, usually with a Google Drive download link and, where available, a sample Xcode project that shows how to wire the model into an app. The collection covers image classification, the long line of YOLO detectors plus DETR-style models, segmentation, super-resolution, depth estimation, image generation, and even TTS, voice conversion, and small language models. Before building your own conversion pipeline, it's worth checking here whether the model you want already exists in CoreML form. The conversions follow the upstream projects' licenses, which also helps you gauge which open-weights models are realistic candidates for a mobile build.