W2GenAI

/

LucidFlux

like 23

Follow

W2GenAI Lab 24

This PR enhances the model card for LucidFlux by:

• Adding pipeline_tag: image-to-image: This improves model discoverability on the Hugging Face Hub for relevant tasks.
• Adding library_name: diffusers: The model is built on Diffusion Transformers and inference code confirms its compatibility with a diffusion pipeline, enabling an automated "how to use" widget.
• Correcting the license tag to mit: The license_link points to a file containing the MIT License terms.
• Updating the paper link: The "Technical Report" link in the header is replaced with a direct link to the official Hugging Face paper page: LucidFlux: Caption-Free Universal Image Restoration via a Large-Scale Diffusion Transformer.
• Adding an explicit Code link: A direct link to the GitHub repository is added in the header for easier access to the source code.

These updates provide clearer, more structured information for users and improve integration with the Hugging Face ecosystem.