diffusers/pr_14839/en/index.md

Diffusers

Diffusers provides pretrained diffusion models and the building blocks for custom image, video, and audio workflows.

It has two main paths.

• DiffusionPipeline supports few-line inference with pretrained checkpoints, plus adapters like LoRA. This is the easy path for generation.
• Modular Diffusers enables composable blocks and ModularPipeline for custom pipelines when you need more control.

Optimizations such as offloading and quantization keep large models runnable on memory-constrained devices. If memory is not an issue, Diffusers also supports torch.compile for faster inference.

Browse trending Diffusers models on the Hub now.

Learn

If you're a beginner, start with the Hugging Face Diffusion Models Course. It covers diffusion theory and how to generate images, fine-tune models, and more with Diffusers.

The Quickstart also includes a copyable agent setup prompt for inference.

Where next

• Inference — load pipelines and run generation
• Optimize and scale — memory, speed, quantization, and serving
• Modular Diffusers — build custom pipelines from blocks
• Train and fine-tune — train diffusion models and adapters
• CLI - run and package pipelines from the command line