🎵 Messy Mashup — Audio Genre Classifier
Upload any audio file (WAV, MP3, FLAC, etc.) and the model will predict its genre. Built with a fine-tuned Audio Spectrogram Transformer (AST) that achieves 0.95 Macro F1 on chaotic multi-song mashups.
The model uses 10-crop Test-Time Augmentation for robust predictions.