Update README.md

README.md

@@ -3,7 +3,16 @@ license: mit
 inference: false
 ---
 
-A simple use case:
+# Introduction
+
+**MERT-v0** is a completely unsupervised model trained on 1000 hour music audios. 
+It is SOTA-comparable on multiple MIR tasks even under probing settings, while keeping fine-tunable on a single 2080Ti. 
+It outperforms Jukebox representation on GTZAN (genre classification) and GiantSteps (key classification) datasets.
+Larger models trained with more data are on the way~
+
+
+
+# Model Usage
 
 ```shell
 from transformers import Wav2Vec2Processor, HubertModel