Apple’s latest AI image editing model, called MGIE, has been released, although it is not yet a complete product. It is now available on GitHub, along with a project paper in PDF format that includes instructions.
The model utilizes text prompts to edit and modify images. It can perform basic image editing tasks such as adjusting contrast or brightness, and altering the white balance. Additionally, it can attempt to interpret more complex commands, such as making a pizza healthier by adding vegetable toppings.
Furthermore, the model has the capability to crop and resize images, autofill image borders, change the appearance of a subject’s hair, eyes, and clothing, and remove objects from an image.
While MGIE is only the initial stage of Apple’s generative AI development, it is unlikely to be integrated into any of Apple’s current devices. However, it certainly indicates what the future may hold for Apple’s AI technology.
