Swift 🧨Diffusers for Mac 1.1 — faster on-device Stable Diffusion with Core ML

AI Impact Summary

Swift 🧨Diffusers for Mac 1.1 introduces on-device diffusion acceleration via Core ML with automatic accelerator selection (GPU, Neural Engine, or CPU) tailored to the host Mac. It expands model coverage by exposing multiple Hugging Face Diffusers models (Stable Diffusion 1.5, 2 Base, 2.1 Base, OFA-Sys/small-stable-diffusion-v0) and runs entirely locally, reducing latency and avoiding online queues. The open-source stack running on macOS (Core ML, PyTorch bridge, Hugging Face Hub) enables downstream teams to integrate, audit, and customize diffusion capabilities while preserving user privacy. Performance varies by hardware, so teams should profile on target Macs to select the best accelerator and model set.