---
title: Qwen2-Audio-7B-Instruct-4bit
canonical_url: "https://www.modelscope.cn/models/mlx-community/Qwen2-Audio-7B-Instruct-4bit"
md_url: "https://www.modelscope.cn/models/mlx-community/Qwen2-Audio-7B-Instruct-4bit.md"
repository: mlx-community/Qwen2-Audio-7B-Instruct-4bit
last_updated: 2026-03-29
license: apache-2.0
model_type:
  - qwen2_audio
architectures:
  - Qwen2AudioForConditionalGeneration
base_model:
  - Qwen/Qwen2-Audio-7B-Instruct
base_model_relation: finetune
parameters: 2.4B
tensor_type:
  - U32
  - BF16
library_name:
  - mlx
  - safetensors
  - pytorch
frameworks:
  - pytorch
inference_backends:
  - "deploy_task vlm/text/emb"
  - "sglang 0.5.2"
  - "vllm 0.9.2"
downloads: 225
stars: 0
tags:
  - mlx
  - mlx-audio
  - qwen2-audio
  - audio
  - speech
  - multimodal
  - 4bit
---

# Qwen2-Audio-7B-Instruct-4bit

> Qwen2-Audio-7B-Instruct-4bit - mlx-community 在 ModelScope 开源的模型。Qwen2-Audio-7B-Instruct (4-bit MLX)

mlx-community/Qwen2-Audio-7B-Instruct-4bit 是 ModelScope 魔搭社区上的 2.4B 参数机器学习模型，采用 apache-2.0 许可，基于 Qwen/Qwen2-Audio-7B-Instruct 构建，可用 deploy_task vlm/text/emb、sglang 0.5.2、vllm 0.9.2 部署。

- **Repository**: mlx-community/Qwen2-Audio-7B-Instruct-4bit
- **License**: apache-2.0
- **Parameters**: 2.4B
- **Base model**: Qwen/Qwen2-Audio-7B-Instruct
- **Inference backends**: deploy_task vlm/text/emb, sglang 0.5.2, vllm 0.9.2
- **Tags**: mlx, mlx-audio, qwen2-audio, audio, speech, multimodal, 4bit
- **Downloads**: 225
- **Stars**: 0
- **Last updated**: 2026-03-29

Source: https://www.modelscope.cn/models/mlx-community/Qwen2-Audio-7B-Instruct-4bit

---

# Qwen2-Audio-7B-Instruct (4-bit MLX)

4-bit quantized version of [Qwen/Qwen2-Audio-7B-Instruct](https://huggingface.co/Qwen/Qwen2-Audio-7B-Instruct) for Apple Silicon via [mlx-audio](https://github.com/Blaizzy/mlx-audio).

## Usage

```python
from mlx_audio.stt.utils import load_model

model = load_model("mlx-community/Qwen2-Audio-7B-Instruct-4bit")

# Transcription
result = model.generate("audio.wav", prompt="Transcribe the audio.")
print(result.text)

# Audio understanding
result = model.generate("audio.wav", prompt="What emotion is the speaker expressing?")
print(result.text)

# Translation
result = model.generate("audio.wav", prompt="Translate the speech to French.")
print(result.text)
```

## Model Details

- **Base model**: Qwen/Qwen2-Audio-7B-Instruct
- **Quantization**: 4-bit (group_size=64), LLM only (encoder and projector kept in bf16)
- **Size**: ~4.2GB (vs ~15GB bf16)
- **Architecture**: Whisper-style encoder (32 layers) + Linear projector + Qwen2-7B LLM

## Capabilities

- Speech transcription (ASR)
- Speech translation
- Audio captioning
- Emotion / sentiment detection
- Environmental sound classification
- Music understanding
- Voice chat (audio-only input)

## Performance

Tested on Apple Silicon (M-series):
- ~4.7 tokens/sec generation (4-bit)
- Accurate transcription matching HuggingFace reference

## Conversion

Converted using mlx-audio with:
- Audio encoder: bf16 (not quantized)
- Multi-modal projector: bf16 (not quantized)
- Language model: 4-bit quantized (group_size=64)