能力标签

🛠

AI工具

神经网络压缩框架

基于 Python · 开源免费，本地部署，数据完全自主可控

英文名：nncf

⭐ 1.2k Stars 🍴 295 Forks 💻 Python 📄 Apache-2.0 🏷 AI 8.0分

8.0AI 综合评分

compressiondeep-learningpython

📺 TG 频道

✦ AI Skill Hub 推荐

经 AI Skill Hub 精选评估，神经网络压缩框架获评「强烈推荐」。已获得 1.2k 颗 GitHub Star，这款AI工具在功能完整性、社区活跃度和易用性方面表现出色，AI 评分 8.0 分，适合有一定技术背景的用户使用。

📚 深度解析

神经网络压缩框架是一款基于 Python 的开源工具，在 GitHub 上收获 1k+ Star，是compression、deep-learning、python领域中的优质开源项目。开源工具的最大优势在于代码完全透明，你可以审计每一行代码的安全性，也可以根据自身需求进行二次开发和定制。

**为什么要使用开源工具而非商业 SaaS？**
对于个人开发者和有隐私需求的用户，本地部署的开源工具意味着数据不离本机，不受第三方服务商的数据政策约束。同时，开源工具通常没有使用次数限制和月度费用，一次安装即可长期使用，对于高频使用场景的总拥有成本（TCO）远低于订阅制商业工具。

**安装与环境准备**
神经网络压缩框架依赖 Python 运行环境。建议通过 pyenv（Python）或 nvm（Node.js）管理 Python 版本，避免全局环境污染。对于新手用户，推荐先创建虚拟环境（python -m venv venv && source venv/bin/activate），再安装依赖，这样即使出现问题也可以随时删除虚拟环境重新开始，不影响系统稳定性。

**社区与维护**
GitHub Issue 和 Discussion 是获取帮助的最快渠道。在提问前建议先检查 Closed Issues（已关闭的问题），大多数常见问题都已有解答。遇到 Bug 时，提供 pip list 的输出、完整错误堆栈和最小可复现示例，能显著提高开发者响应速度。AI Skill Hub 将持续追踪神经网络压缩框架的版本更新，及时通知重要功能变化。

📋 工具概览

神经网络压缩框架是一款基于 Python 开发的开源工具，专注于 compression、deep-learning、python 等核心功能。作为 GitHub 开源项目，它拥有活跃的社区支持和持续的版本迭代，代码完全透明可审计，支持本地部署以保护数据隐私。无论是个人使用还是集成到企业工作流，都能提供稳定可靠的解决方案。

GitHub Stars

⭐ 1.2k

开发语言

Python

支持平台

Windows / macOS / Linux

维护状态

正常维护，社区驱动

开源协议

Apache-2.0

AI 综合评分

8.0 分

工具类型

AI工具

Forks

295

📖 中文文档

以下内容由 AI Skill Hub 根据项目信息自动整理，如需查看完整原始文档请访问底部「原始来源」。

📌 核心特色

开源免费，支持本地部署，数据完全自主可控
活跃的 GitHub 开源社区，持续迭代更新
提供详细文档和使用示例，新手友好
支持自定义配置，灵活适配不同使用环境
可作为基础组件集成进现有技术栈或进行二次开发

🎯 主要使用场景

本地部署运行，保护数据隐私，满足合规要求
自定义集成到现有系统，扩展技术栈能力
作为开源基础组件进行商业化二次开发

以下安装命令基于项目开发语言和类型自动生成，实际以官方 README 为准。

安装命令

# 方式一：pip 安装（推荐）
pip install nncf

# 方式二：虚拟环境安装（推荐生产环境）
python -m venv .venv
source .venv/bin/activate  # Windows: .venv\Scripts\activate
pip install nncf

# 方式三：从源码安装（获取最新功能）
git clone https://github.com/openvinotoolkit/nncf
cd nncf
pip install -e .

# 验证安装
python -c "import nncf; print('安装成功')"

📋 安装步骤说明

访问 GitHub 仓库页面
按照 README 文档完成依赖安装
根据系统环境完成初始化配置
参考官方示例或文档开始使用
遇到问题可在 GitHub Issues 中查找解答

以下用法示例由 AI Skill Hub 整理，涵盖最常见的使用场景。

常用命令 / 代码示例

# 命令行使用
nncf --help

# 基本用法
nncf input_file -o output_file

# Python 代码中调用
import nncf

# 示例
result = nncf.process("input")
print(result)

以下配置示例基于典型使用场景生成，具体参数请参照官方文档调整。

配置示例

# nncf 配置文件示例（config.yml）
app:
  name: "nncf"
  debug: false
  log_level: "INFO"

# 运行时指定配置文件
nncf --config config.yml

# 或通过环境变量配置
export NNCF_API_KEY="your-key"
export NNCF_OUTPUT_DIR="./output"

📑 README 深度解析真实文档完整度 52/100 查看 GitHub 原文 →

以下内容由系统直接从 GitHub README 解析整理，保留代码块、表格与列表结构。

简介

Key Features

Installation Guide

For detailed installation instructions, refer to the Installation guide.

NNCF can be installed as a regular PyPI package via pip:

pip install nncf

NNCF is also available via conda:

conda install -c conda-forge nncf

System requirements of NNCF correspond to the used backend. System requirements for each backend and the matrix of corresponding versions can be found in installation.md.

Usage

Demos, Tutorials and Samples

For a quicker start with NNCF-powered compression, try sample notebooks and scripts presented below.

Jupyter* Notebook Tutorials and Demos

Ready-to-run Jupyter* notebook tutorials and demos are available to explain and display NNCF compression algorithms for optimizing models for inference with the OpenVINO Toolkit:

Notebook Tutorial Name	Compression Algorithm	Backend	Domain
[BERT Quantization](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/language-quantize-bert)<br>[![Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/openvinotoolkit/openvino_notebooks/blob/latest/notebooks/language-quantize-bert/language-quantize-bert.ipynb)	Post-Training Quantization	OpenVINO	NLP
[MONAI Segmentation Model Quantization](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/ct-segmentation-quantize)<br>[![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/openvinotoolkit/openvino_notebooks/HEAD?filepath=notebooks%2Fct-segmentation-quantize%2Fct-scan-live-inference.ipynb)	Post-Training Quantization	OpenVINO	Segmentation
[PyTorch Model Quantization](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/pytorch-post-training-quantization-nncf)	Post-Training Quantization	PyTorch	Image Classification
[YOLOv11 Quantization with Accuracy Control](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/yolov11-quantization-with-accuracy-control)	Post-Training Quantization with Accuracy Control	OpenVINO	Speech-to-Text,<br>Object Detection

A list of notebooks demonstrating OpenVINO conversion and inference together with NNCF compression for models from various domains:

Demo Model	Compression Algorithm	Backend	Domain
[OneFormer](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/oneformer-segmentation)	Post-Training Quantization	OpenVINO	Image Segmentation
[CLIP](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/clip-zero-shot-image-classification)	Post-Training Quantization	OpenVINO	Image-to-Text
[BLIP](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/blip-visual-language-processing)	Post-Training Quantization	OpenVINO	Image-to-Text
[Latent Consistency Model](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/latent-consistency-models-image-generation)	Post-Training Quantization	OpenVINO	Text-to-Image
[Distil-Whisper](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/distil-whisper-asr)	Post-Training Quantization	OpenVINO	Speech-to-Text
[Whisper](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/whisper-subtitles-generation)<br>[![Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/openvinotoolkit/openvino_notebooks/blob/latest/notebooks/whisper-subtitles-generation/whisper-convert.ipynb)	Post-Training Quantization	OpenVINO	Speech-to-Text
[MMS Speech Recognition](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/mms-massively-multilingual-speech)	Post-Training Quantization	OpenVINO	Speech-to-Text
[LLM Instruction Following](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/llm-question-answering)	Weight Compression	OpenVINO	NLP, Instruction Following
[LLM Chat Bots](https://github.com/openvinotoolkit/openvino_notebooks/tree/latest/notebooks/llm-chatbot)	Weight Compression	OpenVINO	NLP, Chat Bot

Post-Training Quantization and Weight Compression Examples

Compact scripts demonstrating quantization/weight compression and corresponding inference speed boost:

Example Name	Compression Algorithm	Backend	Domain
[OpenVINO MobileNetV2](./examples/post_training_quantization/openvino/mobilenet_v2/README.md)	Post-Training Quantization	OpenVINO	Image Classification
[OpenVINO YOLO26](./examples/post_training_quantization/openvino/yolo26/README.md)	Post-Training Quantization	OpenVINO	Object Detection
[OpenVINO YOLOv8 QwAC](./examples/post_training_quantization/openvino/yolov8_quantize_with_accuracy_control/README.md)	Post-Training Quantization with Accuracy Control	OpenVINO	Object Detection
[OpenVINO Anomaly Classification](./examples/post_training_quantization/openvino/anomaly_stfpm_quantize_with_accuracy_control/README.md)	Post-Training Quantization with Accuracy Control	OpenVINO	Anomaly Classification
[PyTorch MobileNetV2](./examples/post_training_quantization/torch/mobilenet_v2/README.md)	Post-Training Quantization	PyTorch	Image Classification
[PyTorch SSD](./examples/post_training_quantization/torch/ssd300_vgg16/README.md)	Post-Training Quantization	PyTorch	Object Detection
[TorchFX Resnet18](./examples/post_training_quantization/torch_fx/resnet18/README.md)	Post-Training Quantization	TorchFX	Image Classification
[ONNX MobileNetV2](./examples/post_training_quantization/onnx/mobilenet_v2/README.md)	Post-Training Quantization	ONNX	Image Classification
[ONNX YOLOv8 QwAC](./examples/post_training_quantization/onnx/yolov8_quantize_with_accuracy_control/README.md)	Post-Training Quantization with Accuracy Control	ONNX	Object Detection
[ONNX TinyLlama WC](./examples/llm_compression/onnx/tiny_llama/README.md)	Weight Compression	ONNX	LLM
[TorchFX TinyLlama WC](./examples/llm_compression/torch_fx/tiny_llama/README.md)	Weight Compression	TorchFX	LLM
[OpenVINO TinyLlama WC](./examples/llm_compression/openvino/tiny_llama/README.md)	Weight Compression	OpenVINO	LLM
[OpenVINO TinyLlama WC with HS](./examples/llm_compression/openvino/tiny_llama_find_hyperparams/README.md)	Weight Compression with Hyperparameters Search	OpenVINO	LLM
[ONNX TinyLlama WC with SE](./examples/llm_compression/onnx/tiny_llama_scale_estimation/README.md)	Weight Compression with Scale Estimation	ONNX	LLM

Quantization-Aware Training Examples

Example Name	Compression Algorithm	Backend	Domain
[PyTorch Resnet18](./examples/quantization_aware_training/torch/resnet18/README.md)	Quantization-Aware Training	PyTorch	Image Classification
[PyTorch Anomalib](./examples/quantization_aware_training/torch/anomalib/README.md)	Quantization-Aware Training	PyTorch	Anomaly Detection

Contributing Guide

Refer to the CONTRIBUTING.md file for guidelines on contributions to the NNCF repository.

Step 3: Run the quantization pipeline

quantized_model = nncf.quantize(model, calibration_dataset)


</details>

<details><summary><b>PyTorch</b></summary>

python import nncf import torch from torchvision import datasets, models

Step 3: Run the quantization pipeline

quantized_model = nncf.quantize(model, calibration_dataset)


**NOTE** If the Post-Training Quantization algorithm does not meet quality requirements you can fine-tune the quantized pytorch model. You can find an example of the Quantization-Aware training pipeline for a pytorch model [here](examples/quantization_aware_training/torch/resnet18/README.md).

</details>

<details><summary><b>TorchFX</b></summary>

python import nncf import torch.fx from torchvision import datasets, models

fx_model = torch.export.export(model, args=(ex_input,)).module()

Step 4: Run the quantization pipeline

quantized_fx_model = nncf.quantize(fx_model, calibration_dataset)


</details>

<details><summary><b>ONNX</b></summary>

python import onnx import nncf import torch from torchvision import datasets

Step 3: Run the quantization pipeline

quantized_model = nncf.quantize(onnx_model, calibration_dataset) ```

</details>

[//]: # (NNCF provides full [samples](#post-training-quantization-samples), which demonstrate Post-Training Quantization usage for PyTorch, ONNX, and OpenVINO.)

Step 3: Run the quantization pipeline

quantized_model = nncf.quantize(model, calibration_dataset)

Now use compressed_model as a usual torch.nn.Module

Save quantization modules and the quantized model parameters

checkpoint = { 'state_dict': model.state_dict(), 'nncf_config': nncf.torch.get_config(model), ... # the rest of the user-defined objects to save } torch.save(checkpoint, path_to_checkpoint)