luklebigbosse

/

BM-MAE

like 1

This PR significantly enhances the model card by:

• Adding the pipeline_tag: image-to-image to the metadata, making the model discoverable under relevant tasks on the Hub, particularly given its ability to reconstruct missing modalities.
• Including the comprehensive abstract from the paper for better understanding of the model's context and capabilities.
• Integrating the model's framework image for visual clarity.
• Providing a "Quickstart" section with a Python code snippet from the original GitHub repository, demonstrating how to extract features using the ViTEncoder.
• Adding a code example for loading the model directly from the Hugging Face Hub using from_pretrained, as suggested in the project's GitHub README.
• Adding the BibTeX citation for proper academic attribution.