Hugging Face releases Modular Diffusers framework
hugging face launched modular diffusers, a framework breaking diffusion pipelines into composable blocks. the system allows custom block creation, independent execution of components like text encoding or denoising, and integrates with the visual interface mellon. the community released a 14b parameter real-time video model and a 2.3b world model using the new architecture.
Hugging Face released Modular Diffusers, a framework that decouples diffusion pipelines into flexible, composable blocks for text encoding, image encoding, denoising, and decoding. Users can inspect, run, or swap blocks independently via `.init_pipeline()` and share custom blocks via the Hub using a new `ModularPipeline` class. The release includes integration with Mellon, a node-based visual workflow editor.
Early adopters have released complex pipelines using the framework, including 'Krea Realtime Video' (a 14B parameter model distilled from Wan 2.1 achieving 11fps) and 'Waypoint-1' (a 2.3B parameter real-time world model).