Starstruk Logo MarkStarstruk

v2.0

RyanCodrai/turbovecv2.0Mar 31, 2026by lipku

AI Summary

A major refactoring that consolidates common logic into base classes, switches models to a plugin architecture, and adds RTMP output support.

Key Highlights

  • Code refactoring: Moved digital avatar and audio feature code to the `avatars` directory
  • Architecture change: Models and TTS now use plugin-based integration
  • Stream output logic moved to `streamout` directory with new RTMP output support
  • Common inference and response logic consolidated into `BaseAvatar`
  • Audio feature splitting logic consolidated into `BaseAsr`
  • Added Aliyun Qwen TTS support

Breaking Changes

  • Code refactoring and architecture changes to plugin-based models and directory structure

New Features

  • RTMP output support
  • Aliyun Qwen TTS integration

Full Release Notes

1，重构代码，数字人模型和音频特征代码移到avatars目录下
2，数字人模型、tts、传输方式改成plugin方式接入
3，传输方式整理成单独的类，放到streamout目录下，添加rtmp输出
4，数字人推理和回贴公共部分代码整合到BaseAvatar中，各模型只需要实现自己特有部分代码
5，音频特征的切分代码整合到BaseAsr中
6，tts添加阿里云qwentts