deepinsight/insightface ? reverse-engineered prompt

I want to build a powerful Python toolkit for advanced 2D and 3D face analysis. The main functions should be face detection to locate faces in an image, face recognition to identify specific people, and face alignment to map out their facial landmarks. It should also support more advanced features like estimating a person's age and even reconstructing a 3D model of a face from a single 2D photo.

A key feature for me is face swapping. I want the ability to easily take a face from a source image and swap it onto a person in a target image or even a video.

Please use Python and PyTorch for this project. It should be designed to use state-of-the-art pre-trained models for all these tasks, so I don't have to do any training myself. It would also be great if you could include a few simple web demos to showcase the main capabilities like detection, recognition, and the face swapping feature.