AI GitHub
Diffusers for building generation pipelines in Python
Use modular diffusion pipelines for supported image, video and audio models.
Key Takeaways
- A pipeline connects model components, a scheduler and input processing.
- Choose the pipeline and model components from the checkpoint's documentation.
- Start with a documented pipeline and fixed generator seed.
- Pipelines vary in their inputs and artifact requirements.
How it works
A pipeline connects model components, a scheduler and input processing. Python code can replace individual parts or integrate generation into an application.
What you need
Choose the pipeline and model components from the checkpoint's documentation. Keep preprocessing and scheduler choices with the generated image when recording a result.
What to try first
Start with a documented pipeline and fixed generator seed. Compare two scheduler settings with the same prompt, then record output and memory use.
Log the checkpoint revision, scheduler configuration and preprocessing. A change to any of them can cause a different image even if the prompt and seed are fixed.
Where to look for failures
Pipelines vary in their inputs and artifact requirements. Check the specific model card and dependency versions before expecting a workflow to transfer.
Related Franklin pages
Match the pipeline and device to a specific model variant. Compare a controlled prompt on known assets before testing a new checkpoint, conditioning input or export.
- Guide supported image generation models with inputs such as edges, poses or depth. ControlNet for conditioning diffusion with spatial cues
- Explore Black Forest Labs' official inference code and model references. FLUX reference code for image generation
Franklin’s image tests give you examples of outputs to compare across visual models. Franklin’s image tests
Project documentation
Try it for your workflow
Explore the project’s documentation for setup and usage, or save this page to a collection for later.
Read the project documentation ↗