A framework that optimizes AI models to run efficiently on Apple Silicon chips (like M1, M2, M3), taking advantage of their specific hardware capabilities.