Apple Silicon local inference

Running models on Apple chips for fast, private processing without cloud calls. See tips on setup, optimization, and on-device AI workloads here.