Multi Model Runtime Selection And Hot Swapping

via “multi-model-management-and-switching”

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 a message-based model state machine (mltl=model loading started, mlpr=model loading progress, mdld=model loaded) that keeps the frontend responsive during long-running model operations. The backend uses PyTorch's model.to(device) and del operations to explicitly manage VRAM, avoiding garbage collection delays.

vs others: More user-friendly than command-line model management (no manual environment setup) and faster than running separate Python processes for each model, while providing better memory efficiency than keeping all models loaded simultaneously.