This PR improves the model card for AudioStory by adding key metadata and essential links.
Changes include:
- Adding `license: apache-2.0` based on the repository's license information.
- Adding `pipeline_tag: text-to-audio` as the model is designed for text-to-audio generation.
- Adding `library_name: transformers` as evidence from the `config.json` and `tokenizer_config.json` files indicates compatibility with the Hugging Face Transformers library (e.g., Qwen2 and T5 architectures, Qwen2Tokenizer class).
- Adding a direct link to the Hugging Face paper page: [AudioStory: Generating Long-Form Narrative Audio with Large Language Models](https://huggingface.co/papers/2508.20088).
- Updating relative image paths to absolute GitHub URLs for `audiostory.png` and `audiostory_framework.png` to ensure proper display on the Hub.
These enhancements will make the model more discoverable and easier to understand for users on the Hugging Face Hub.
