Add model card with pipeline_tag to enable serverless inference

README.md

@@ -0,0 +1,32 @@
+---
+pipeline_tag: audio-classification
+library_name: transformers
+language: en
+tags:
+  - audio
+  - audio-classification
+  - ast
+  - music
+  - mood
+license: apache-2.0
+---
+
+# Audio Mood Classifier
+
+Fine-tuned [Audio Spectrogram Transformer (AST)](https://huggingface.co/MIT/ast-finetuned-audioset-10-10-0.4593) for music mood classification.
+
+## Classes
+
+| Label | Description |
+|---|---|
+| calm_melancholic | Slow, introspective, melancholic |
+| moderate_neutral | Balanced, mid-energy, neutral feel |
+| energetic_upbeat | Fast, high-energy, upbeat |
+
+## Usage
+
+```python
+from transformers import pipeline
+classifier = pipeline("audio-classification", model="guyPerry/audio-mood-classifier")
+result = classifier("song.mp3")
+```