WestlakeAudioLab

/

CleanMel

like 0

Follow

Audio Lab of the Westlake University 4

This PR significantly improves the model card for the CleanMel model by:

• Standardizing the pipeline_tag to audio-to-audio for accurate categorization on the Hugging Face Hub.
• Adding a comprehensive overview of the model based on its abstract.
• Including direct links to the official Hugging Face paper page, project page, GitHub repository, and Hugging Face Space demo for easy access to resources.
• Integrating a "Quick Start" section with detailed environment setup and correct inference commands from the GitHub repository.
• Featuring performance metrics and visual results for both speech enhancement and ASR accuracy.
• Adding the full BibTeX citation and acknowledgments.

This update makes the model card more informative and user-friendly, enhancing discoverability and usability for the community.