Add model card

This PR adds a model card, improving discoverability by:

- linking the model to the paper at https://huggingface.co/papers/2412.15322.
- making the model discoverable at https://huggingface.co/models?pipeline_tag=audio-to-audio.
- specifying the library and license.

README.md

@@ -1,8 +1,10 @@
+```yaml
 ---
 license: mit
-language:
-- en
+library_name: diffusers
+pipeline_tag: audio-to-audio
 ---
-Code: https://github.com/hkchengrex/MMAudio
-license: mit
----
+This repository contains the model described in [Taming Multimodal Joint Training for High-Quality Video-to-Audio Synthesis](https://huggingface.co/papers/2412.15322).
+
+Code: https://github.com/hkchengrex/MMAudio.
+```