MusicLM
lastmod 2023-02-01created 2023-02-013 edits
AI that generates music from text
Features
- Paper published by Google Research
- Generates music highly faithful to the text description, e.g. "a calming violin melody backed by a distorted guitar riff"
- Casts it as a hierarchical sequence-to-sequence modeling task to generate several minutes of 24kHz music
- Far outperforms existing systems in both audio quality and adherence to the text description
- Can also transform melodies in the form of whistling or humming
- Can generate audio in story mode by providing multiple prompts sequentially
- Released the MusicCaps dataset of 5.5k music-text pairs
See also
- Deep learning
- Magenta
- DeepJazz
- BachBot
- FlowMachines
- WaveNet
- GRUV
- MuseNet
- MusicLM - AI that generates music from text
- Riffusion - Stable Diffusion fine-tuned to generate music