---
title: README
emoji: ⚡
colorFrom: green
colorTo: gray
sdk: static
pinned: false
---

# Audio, Music, and AI Lab (AMAAI)

The Audio, Music, and AI lab at Singapore University of Technology and Design focuses on cutting-edge innovation in Audio and Music AI. 

[More info and publications here.](https://dorienherremans.com/biblio) 

Popular software: 
- Mustango: text-to-music generation: [code](https://github.com/AMAAI-Lab/mustango) - [live demo](https://replicate.com/declare-lab/mustango)
- Video2Music: video-to-music generation: [code](https://github.com/AMAAI-Lab/Video2Music)
- Music2Emo: emotion detection from music: [code](https://github.com/AMAAI-Lab/Music2Emotion) - [live demo](https://huggingface.co/spaces/amaai-lab/music2emo/discussions/1)
- Text2midi: text-to-midi generation: [code](https://github.com/AMAAI-Lab/text2midi) 
- nnAudio: on-the-fly spectrogram extraction: [code](https://github.com/AMAAI-Lab/nnAudio)

Popular Datasets: 
- [JamendoMaxCaps](https://huggingface.co/datasets/amaai-lab/JamendoMaxCaps): text captions with instrumental music audio
- [MusicBench](https://huggingface.co/datasets/amaai-lab/MusicBench): text captions with music audio
- [MidiCaps](https://huggingface.co/datasets/amaai-lab/MidiCaps): text captions with music midi (large-scale)