This repo is the same as Wombleyy's repo
(1) Supports customized phoneme
[Kokoro](/kˈOkəɹO/) is an open-weight TTS model with 82 million parameters
(2) Supports multi-model with weight modification
docker pull astroneko404/kokoro-offline:latest
docker run -p 8000:8000 astroneko404/kokoro-offline:latest