This PR improves the CCFQA dataset card by enhancing its discoverability and utility.
Specifically, it:
- Adds `audio-text-to-text` to `task_categories` to accurately reflect its cross-modal nature for speech and text factuality evaluation.
- Includes `library_name: datasets` for direct integration into the Hugging Face `datasets` library.
- Adds comprehensive `tags` such as `multilingual`, `cross-modal`, `speech`, `text`, `factuality`, and `evaluation` to provide richer context and improve searchability.
The paper link remains the existing arXiv link as per guidelines.
