Improve model card for SimVQ: Add pipeline tag, paper, GitHub, and usage instructions
#5
by
nielsr
HF Staff
- opened
This PR significantly improves the model card for the SimVQ model by:
- Adding the
pipeline_tag: audio-to-audio, which helps users discover the model for relevant tasks on the Hugging Face Hub, in line with its application to audio processing (e.g., LibriTTS). - Linking the model to its official paper: Addressing Representation Collapse in Vector Quantized Models with One Linear Layer.
- Providing a direct link to the GitHub repository for easy access to the code.
- Including comprehensive usage instructions, including installation, training, and evaluation scripts, directly extracted from the project's GitHub README, ensuring accuracy and utility for users.
- Incorporating the paper's abstract, quantitative comparison tables, and reconstruction visualizations to provide a detailed overview of the model and its performance.
This update ensures the model card is comprehensive, discoverable, and accurately reflects the model's details and usage instructions.