Voice Generation: CosyVoice Deployment

更新时间:
复制 MD 格式

CosyVoice is an open-source voice generation model from Alibaba Cloud that produces high-quality speech synthesis in multiple languages and voices. PAI-EAS provides two deployment options: Standard Edition and High-performance Edition.

Choose the edition that suits your use case.

Edition

Description

Guide

Standard Edition

The Standard Edition is for single-machine testing and applications. It supports both WebUI and API calls.

Quickly deploy a WebUI service

High-performance Edition

Uses a separate frontend and backend architecture. A single backend instance can handle traffic from up to eight frontend instances, which is ideal for large-scale business scenarios. It also supports both WebUI and API calls.

Quickly deploy a high-performance service with a separate frontend and backend