farzadbz

/

Medical-VAE

like 1

This PR enhances the model card for the farzadbz/Medical-VAE models by:

• Adding the library_name: diffusers metadata tag for improved discoverability and compatibility with the diffusers ecosystem, as these VAEs are based on a latent diffusion architecture.
• Clarifying their role as foundational components for both the recently published REFLECT: Rectified Flows for Efficient Brain Anomaly Correction Transport paper and the previously cited MAD-AD: Masked Diffusion for Unsupervised Brain Anomaly Detection.
• Including relevant GitHub repository links for both REFLECT and MAD-AD.
• Updating the "Usage" section to include .sample() on the decode output, which is common practice for VAEs that return a distribution.
• Adding the BibTeX citation for the REFLECT paper to the "Citation & Reference" section.