Update README.md

README.md

@@ -9,7 +9,8 @@ library_name: nnAudio
 license: mit
 ---
 
-# Myna: Masking-Based Contrastive Learning of Musical Representations
+# Myna-Base
+## From "Myna: Masking-Based Contrastive Learning of Musical Representations"
 
 ## Model Overview
 Myna is a self-supervised contrastive model designed for musical representation learning. It employs a Vision Transformer (ViT) backbone on mel-spectrograms and introduces token masking as its primary augmentation method. Unlike traditional contrastive learning frameworks that rely on augmentations such as pitch shifts, Myna retains pitch sensitivity, leading to improvements in key detection tasks.