---
title: Qwen3-1.7B-dwq-3bit-gs128
canonical_url: "https://www.modelscope.cn/models/mlx-community/Qwen3-1.7B-dwq-3bit-gs128"
md_url: "https://www.modelscope.cn/models/mlx-community/Qwen3-1.7B-dwq-3bit-gs128.md"
repository: mlx-community/Qwen3-1.7B-dwq-3bit-gs128
last_updated: 2025-07-17
license: mit
model_type:
  - qwen3
architectures:
  - Qwen3ForCausalLM
library_name:
  - mlx
  - pytorch
frameworks:
  - pytorch
inference_backends:
  - "deploy_task text/emb"
  - "lmdeploy_turbomind 0.9.1"
  - "sglang 0.5.2"
  - "vllm 0.9.2"
downloads: 157
stars: 0
---

# Qwen3-1.7B-dwq-3bit-gs128

> Qwen3-1.7B-dwq-3bit-gs128 - mlx-community 在 ModelScope 开源的模型。Qwen3-1.7B-dwq-3bit-gs128

mlx-community/Qwen3-1.7B-dwq-3bit-gs128 是 ModelScope 魔搭社区上的机器学习模型，采用 mit 许可，可用 deploy_task text/emb、lmdeploy_turbomind 0.9.1、sglang 0.5.2 部署。

- **Repository**: mlx-community/Qwen3-1.7B-dwq-3bit-gs128
- **License**: mit
- **Inference backends**: deploy_task text/emb, lmdeploy_turbomind 0.9.1, sglang 0.5.2, vllm 0.9.2
- **Downloads**: 157
- **Stars**: 0
- **Last updated**: 2025-07-17

Source: https://www.modelscope.cn/models/mlx-community/Qwen3-1.7B-dwq-3bit-gs128

---

# Qwen3-1.7B-dwq-3bit-gs128

This model is a 3-bit DWQ quantized version of [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B) using MLX.

## Quantization Details

- **Method**: DWQ (Dynamic Weight Quantization)
- **Bits**: 3
- **Group Size**: 128
- **Quantized Layers**: 197

## Usage

```python
import mlx.core as mx
from mlx_lm import load, generate

model, tokenizer = load("mlx-community/Qwen3-1.7B-dwq-3bit-gs128")
response = generate(model, tokenizer, prompt="Hello, world!", verbose=True)
```

## Original Model

This is a quantized version of [Qwen/Qwen3-1.7B](https://huggingface.co/Qwen/Qwen3-1.7B).