Macos Native Inference With Mlx Framework Acceleration

via “apple-silicon-metal-acceleration-for-inference”

Diffusion Bee is the easiest way to run Stable Diffusion locally on your M1 Mac. Comes with a one-click installer. No dependencies or technical knowledge needed.

Unique: Implements runtime processor detection and conditional PyTorch backend selection, automatically using Metal Performance Shaders on Apple Silicon while gracefully falling back to CPU on Intel Macs. The system profiles operation performance and selectively offloads to Metal only for operations where it provides speedup.

vs others: Faster than CPU-only inference (3-5x speedup on M1/M2) and more accessible than CUDA-based acceleration (no NVIDIA GPU required), while maintaining compatibility with Intel Macs through automatic fallback.