---
title: Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16
canonical_url: "https://www.modelscope.cn/models/okwinds/Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16"
md_url: "https://www.modelscope.cn/models/okwinds/Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16.md"
repository: okwinds/Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16
chinese_name: "通义千问 Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16 量化高精校准"
last_updated: 2025-08-07
license: apache-2.0
pipeline_tag: text-generation
tasks:
  - text-generation
model_type:
  - qwen3_moe
architectures:
  - Qwen3MoeForCausalLM
base_model:
  - Qwen/Qwen3-Coder-30B-A3B-Instruct
base_model_relation: quantized
parameters: 8.1B
tensor_type:
  - BF16
  - I64
  - I32
library_name:
  - transformer
  - safetensors
  - pytorch
frameworks:
  - Pytorch
language:
  - en
  - zh
inference_backends:
  - "deploy_task text/emb"
  - "lmdeploy_turbomind 0.9.1"
  - "sglang 0.5.2"
  - "vllm 0.9.2"
downloads: 1138
stars: 5
tags:
  - Thinking
  - Chat
  - "Reason Model"
  - Quantization
---

# Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16

> Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16 - okwinds 在 ModelScope 开源的模型。重要：友情提醒，推理本模型时，请按照如下官方最佳实践：

okwinds/Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16 是 ModelScope 魔搭社区上的 8.1B 参数text-generation模型，采用 apache-2.0 许可，基于 Qwen/Qwen3-Coder-30B-A3B-Instruct 构建，可用 deploy_task text/emb、lmdeploy_turbomind 0.9.1、sglang 0.5.2 部署。

- **Repository**: okwinds/Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16
- **License**: apache-2.0
- **Tasks**: text-generation
- **Parameters**: 8.1B
- **Base model**: Qwen/Qwen3-Coder-30B-A3B-Instruct
- **Inference backends**: deploy_task text/emb, lmdeploy_turbomind 0.9.1, sglang 0.5.2, vllm 0.9.2
- **Tags**: Thinking, Chat, Reason Model, Quantization
- **Downloads**: 1138
- **Stars**: 5
- **Last updated**: 2025-08-07

Source: https://www.modelscope.cn/models/okwinds/Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16

---

### <span style="color:red">重要：友情提醒，推理本模型时，请按照如下官方最佳实践：</span>

为了达到最佳性能，建议以下设置：

1. **采样参数**:
   - 建议使用 `Temperature=0.7`, `TopP=0.8`, `TopK=20`, 以及 `presence_penalty=1.05`.

2. **足够的输出长度**: 对于大多数查询，建议使用 65,536 个 tokens 的输出长度，这对于指令模型是足够的。


#### 📖 关于量化损失方面的研究，可阅读公众号“觉察流”文章👇</br>

《[Reason Model 的“瘦身计划”：量化技术的得与失](https://mp.weixin.qq.com/s/NMGq4UUkfo8GMix5LHnWCg)》


# 通义千问 Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16 量化高精校准

原模型 [Qwen/Qwen3-Coder-30B-A3B-Instruct](https://www.modelscope.cn/models/Qwen/Qwen3-Coder-30B-A3B-Instruct)

---

#### _作者在此 👇🏻 扫一扫_

<img src="https://www.modelscope.cn/models/okwinds/GPT-2/resolve/master/qrcode_for_jcl_258.jpg" />

---

## 下载

SDK下载
```bash
#安装ModelScope
pip install modelscope
```
```python
#SDK模型下载
from modelscope import snapshot_download
model_dir = snapshot_download('okwinds/Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16')
```
Git下载
```
#Git模型下载
git clone https://www.modelscope.cn/okwinds/Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16.git
```

## 模型概述

Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16 是一个基于 通义千问 Qwen3-Coder-30B-A3B-Instruct 的 INT8 量化并校准的模型。

- **模型名称:** Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16
- **模型架构:** Qwen3 MoE
- **权重量化:** INT8

该模型通过将 [qwen/Qwen3-Coder-30B-A3B-Instruct](https://www.modelscope.cn/models/Qwen/Qwen3-Coder-30B-A3B-Instruct) 的权重量化为 INT8 数据类型而实现。

量化过程将每个参数从 16bit 减少到 8bit，将模型占用磁盘空间大小，以及推理时加载模型需要的GPU显存空间，减少到了大约为原模型的1/2。

量化过程中，只有 transformer 中的 Linear 层权重是量化的，其他层均保持为原模型数据类型，采用混合精度的计算方式，尽可能减少模型量化后所带来的精度损失。

在量化过程中，使用了[AutoGPTQ](https://github.com/AutoGPTQ/AutoGPTQ)，采用 Symmetric group-wise 方式量化。并在此次量化过程中，进行了数据校准，以提升模型生成精度。（相对比 BF16 做到几乎无损）

> <span style="color: red;">注意：本模型需要 compute capability > 8.0（Ampere、Ada Lovelace、Hopper 架构）的 Nvidia GPU 来支持 INT8 **混合精度**计算。</span>

## 部署推理

#### 推荐使用 vLLM>=0.8.4 ( transformers>=4.51.0 )

Openai api 兼容模式

```bash
>>> vllm serve "/home/gavin/llm/Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16" --host 0.0.0.0 --port 8000 --gpu-memory-utilization 0.9 --served-model-name "Qwen3-Coder-30B-A3B-Instruct-Int8-W8A16"
```

# 附录  Qwen3-Coder-30B-A3B-Instruct 介绍

<a href="https://chat.qwen.ai/" target="_blank" style="margin: 2px;">
    <img alt="Chat" src="https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5" style="display: inline-block; vertical-align: middle;"/>
</a>

## Highlights

**Qwen3-Coder** is available in multiple sizes. Today, we're excited to introduce **Qwen3-Coder-30B-A3B-Instruct**. This streamlined model maintains impressive performance and efficiency, featuring the following key enhancements:  

- **Significant Performance** among open models on **Agentic Coding**, **Agentic Browser-Use**, and other foundational coding tasks.
- **Long-context Capabilities** with native support for **256K** tokens, extendable up to **1M** tokens using Yarn, optimized for repository-scale understanding.
- **Agentic Coding** supporting for most platform such as **Qwen Code**, **CLINE**, featuring a specially designed function call format.

![image/jpeg](https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen3-Coder/qwen3-coder-30a3-main.jpg)

## Model Overview

**Qwen3-Coder-30B-A3B-Instruct** has the following features:
- Type: Causal Language Models
- Training Stage: Pretraining & Post-training
- Number of Parameters: 30.5B in total and 3.3B activated
- Number of Layers: 48
- Number of Attention Heads (GQA): 32 for Q and 4 for KV
- Number of Experts: 128
- Number of Activated Experts: 8
- Context Length: **262,144 natively**. 

**NOTE: This model supports only non-thinking mode and does not generate ``<think></think>`` blocks in its output. Meanwhile, specifying `enable_thinking=False` is no longer required.**

For more details, including benchmark evaluation, hardware requirements, and inference performance, please refer to our [blog](https://qwenlm.github.io/blog/qwen3-coder/), [GitHub](https://github.com/QwenLM/Qwen3-Coder), and [Documentation](https://qwen.readthedocs.io/en/latest/).


## Quickstart

We advise you to use the latest version of `transformers`.

With `transformers<4.51.0`, you will encounter the following error:
```
KeyError: 'qwen3_moe'
```

The following contains a code snippet illustrating how to use the model generate content based on given inputs. 
```python
from transformers import AutoModelForCausalLM, AutoTokenizer

model_name = "Qwen/Qwen3-Coder-30B-A3B-Instruct"

# load the tokenizer and the model
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="auto",
    device_map="auto"
)

# prepare the model input
prompt = "Write a quick sort algorithm."
messages = [
    {"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)

# conduct text completion
generated_ids = model.generate(
    **model_inputs,
    max_new_tokens=65536
)
output_ids = generated_ids[0][len(model_inputs.input_ids[0]):].tolist() 

content = tokenizer.decode(output_ids, skip_special_tokens=True)

print("content:", content)
```

**Note: If you encounter out-of-memory (OOM) issues, consider reducing the context length to a shorter value, such as `32,768`.**

For local use, applications such as Ollama, LMStudio, MLX-LM, llama.cpp, and KTransformers have also supported Qwen3.

## Agentic Coding

Qwen3-Coder excels in tool calling capabilities. 

You can simply define or use any tools as following example.
```python
# Your tool implementation
def square_the_number(num: float) -> dict:
    return num ** 2

# Define Tools
tools=[
    {
        "type":"function",
        "function":{
            "name": "square_the_number",
            "description": "output the square of the number.",
            "parameters": {
                "type": "object",
                "required": ["input_num"],
                "properties": {
                    'input_num': {
                        'type': 'number', 
                        'description': 'input_num is a number that will be squared'
                        }
                },
            }
        }
    }
]

import OpenAI
# Define LLM
client = OpenAI(
    # Use a custom endpoint compatible with OpenAI API
    base_url='http://localhost:8000/v1',  # api_base
    api_key="EMPTY"
)
 
messages = [{'role': 'user', 'content': 'square the number 1024'}]

completion = client.chat.completions.create(
    messages=messages,
    model="Qwen3-Coder-30B-A3B-Instruct",
    max_tokens=65536,
    tools=tools,
)

print(completion.choice[0])
```

## Best Practices

To achieve optimal performance, we recommend the following settings:

1. **Sampling Parameters**:
   - We suggest using `temperature=0.7`, `top_p=0.8`, `top_k=20`, `repetition_penalty=1.05`.

2. **Adequate Output Length**: We recommend using an output length of 65,536 tokens for most queries, which is adequate for instruct models.


### Citation

If you find our work helpful, feel free to give us a cite.

```
@misc{qwen3technicalreport,
      title={Qwen3 Technical Report}, 
      author={Qwen Team},
      year={2025},
      eprint={2505.09388},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2505.09388}, 
}
```