---
title: QwenAudio-Toolkits
canonical_url: "https://www.modelscope.cn/models/funaudio_public/QwenAudio-Toolkits"
md_url: "https://www.modelscope.cn/models/funaudio_public/QwenAudio-Toolkits.md"
repository: funaudio_public/QwenAudio-Toolkits
last_updated: 2026-09-20
license: "Apache License 2.0"
pipeline_tag: auto-speech-recognition
tasks:
  - auto-speech-recognition
library_name:
  - onnx
  - gguf
downloads: 159
stars: 0
tags:
  - audio
  - speech
  - asr
  - tts
  - gguf
---

# QwenAudio-Toolkits

> QwenAudio-Toolkits - funaudio_public 在 ModelScope 开源的模型。QwenAudio Toolkits Model Repository

funaudio_public/QwenAudio-Toolkits 是 ModelScope 魔搭社区上的auto-speech-recognition模型，采用 Apache License 2.0 许可。

- **Repository**: funaudio_public/QwenAudio-Toolkits
- **License**: Apache License 2.0
- **Tasks**: auto-speech-recognition
- **Tags**: audio, speech, asr, tts, gguf
- **Downloads**: 159
- **Stars**: 0
- **Last updated**: 2026-09-20

Source: https://www.modelscope.cn/models/funaudio_public/QwenAudio-Toolkits

---

# QwenAudio Toolkits Model Repository

This repository hosts downloadable model assets for QwenAudio Toolkits. Model
weights are not bundled with the desktop application and are downloaded only
when a user installs a model.

## Layout

```text
models/
  <plugin-id>/
    <variant-id>/
      ... model weights and auxiliary files

runtimes/
  <versioned-runtime-package>/
    <platform>/
      ... native executables and libraries
```

Compatible models share a versioned runtime package. Runtime package IDs are
immutable: an ABI or dependency update is published under a new package ID.

Current platform names are `macos-arm64`, `macos-x64`, `windows-x64`,
`linux-x64`, `linux-arm64`, and `linux-armv7`.

The `plugin.json` files retain model metadata and upstream source information.
The released model catalog determines which entries are shown by the app.

## Converted models

ZipEnhancer and MossFormer2 are mirrored as portable ONNX graphs so the desktop
application can run them without Python or PyTorch. Their upstream links,
conversion notes, licenses, and validation results are recorded next to each
plugin. ONNX Runtime 1.27.0 is stored as a shared runtime package; the current
desktop release uses the tested `macos-arm64` build.
