The Docker deployment steps are as follows:
- Installing Docker and Docker Compose
- Cloning Warehouse:
git clone https://github.com/devnen/Kitten-TTS-Server.git - Select the boot command based on the hardware:
GPU Version::docker compose up -d --build
CPU version::docker compose -f docker-compose-cpu.yml up -d --build - interviews
http://localhost:8005Access to services
Key Notes:
- GPU environment needs to be pre-installedNVIDIA Container Toolkit
- Default port 8005 can be changed in case of conflict
config.yamlhit the nail on the headserver.port - View Log:
docker compose logs -f - Model files are only downloaded on first startup (approx. 25MB)
Docker solutions provide better isolation of dependent environments than native installations.
This answer comes from the articleKitten-TTS-Server: a self-deployable lightweight text-to-speech serviceThe
May not be reproduced without permission:AI productivity tools " How to deploy Kitten-TTS-Server via Docker and what are the considerations?
English 
简体中文 
日本語 
Deutsch 
Português do Brasil