Add pipeline tag

This PR ensures the model can be found at https://huggingface.co/models?pipeline_tag=audio-to-audio&sort=trending

README.md

@@ -4,6 +4,7 @@ datasets:
 language:
 - en
 license: apache-2.0
+pipeline_tag: audio-to-audio
 tags:
 - sound language model
 ---