# Handler for AudioLDM2
This repo can be used to deploy an endpoint for creating music with `AudioLDM2`. It implements most of the optimizations from https://huggingface.co/blog/audioldm2.

The following parameters are exposed:
- inputs (the description of the song)
- duration (the length, defaults to 30s)
- negative prompt (the negative propmt, defaults to `Low quality, average quality.`)
---
license: mit
---