Bark
lastmod 2023-04-25created 2023-04-251 edits
A Text-To-Audio model based on Korean: 트랜스포머
Features
- Can generate highly realistic multilingual speech
- Includes music, background noise, and simple sound effects
- Can also produce nonverbal communication such as laughing, sighing, and crying
- Supports English/Korean and German/French/Japanese/Chinese, etc.
- Uses Facebook's EnCodec codec
- Bark is under the CC-BY 4.0 NC license, and the Suno model itself can be used for commercial purposes