Apple’s AI model for image editing has arrived, though it’s still far from a finished product. MGIE, which stands for Multimodal Large Language Model Guided Image Editing, is currently available on GitHub and has a PDF project paper with instructions.
This model uses text instructions to change and edit images. It can perform simple image editing tasks such as increasing contrast or brightness and changing white balance. But it can also try and explain more complex tips – like making a pizza healthier by adding vegetable toppings.
The model can also crop and resize, automatically fill image borders, change the subject’s hair, eyes, and clothing, and delete objects.
MGIE is just the first step in Apple’s journey to generative AI. It’s unlikely to appear on any of Apple’s current devices, but it could be a sign of things to come.
Source | via