This repository demonstrates the Latent Consistency Model (LCM) and its LoRA variants using the Diffusers library. It provides multiple pipelines including text-to-image, image-to-image, and ControlNet-guided generation. The app runs a local server with MJPEG streaming for real-time webcam-based generation, supporting as few as 4 inference steps via LCM-LoRA acceleration. It supports both SD 1.5 and SDXL models.