Github 项目 - 深度网络可视化工具 - Netron

Author： AIHGF
发布时间：January 17, 2019
2142 views
No comments
3873 words
Categories：网络模型深度平台

> Github 项目 - [Netron](https://github.com/lutzroeder/Netron)
>
> 作者：[lutzroeder](https://github.com/lutzroeder)
>
> 官网：https://www.lutzroeder.com/ai/

![image](https://www.aiuai.cn/uploads/sina/5ce8e0b115b68.jpg)

Netron 是一个强大的神经网络、深度学习和机器学习模型可视化工具. 其支持：

- **ONNX** (`.onnx`, `.pb`, `.pbtxt`)

- **Keras** (`.h5`, `.keras`)
- **CoreML** (`.mlmodel`)
- **Caffe2** (`predict_net.pb`, `predict_net.pbtxt`)
- **MXNet** (`.model`, `-symbol.json`)
- **TensorFlow Lite** (`.tflite`)

Netron 已经测试支持的深度学习框架有：

- **Caffe** (`.caffemodel`, `.prototxt`)

- **PyTorch** (`.pth`)

- **Torch** (`.t7`)

- **CNTK** (`.model`, `.cntk`)

- **PaddlePaddle**(`__model__`)

- **Darknet** (`.cfg`)

- **scikit-learn** (`.pkl`)

- **TensorFlow.js** (`model.json`, `.pb`)

- **TensorFlow** (`.pb`, `.meta`, `.pbtxt`)

可视化效果真心赞！

![image](https://www.aiuai.cn/uploads/sina/5ce8e0b17c8f7.jpg)

## 1. Netron 安装

**macOS**: [**Download**](https://github.com/lutzroeder/netron/releases/latest) the `.dmg` file or run `brew cask install netron`

**Linux**: [**Download**](https://github.com/lutzroeder/netron/releases/latest) the `.AppImage` or `.deb` file.

**Windows**: [**Download**](https://github.com/lutzroeder/netron/releases/latest) the `.exe` installer.

**Browser**: [**Start**](https://www.lutzroeder.com/ai/netron) the browser version.

**Python Server**: Run `pip install netron` and `netron -b [MODEL_FILE]`. In Python run `import netron` and `netron.start('model.onnx')`.

## 2. 下载开源模型

**ONNX Models**: [Inception v1](https://s3.amazonaws.com/download.onnx/models/inception_v1.tar.gz), [Inception v2](https://s3.amazonaws.com/download.onnx/models/inception_v2.tar.gz), [ResNet-50](https://s3.amazonaws.com/download.onnx/models/resnet50.tar.gz), [SqueezeNet](https://s3.amazonaws.com/download.onnx/models/squeezenet.tar.gz)

**Keras Models**: [resnet](https://github.com/Hyperparticle/one-pixel-attack-keras/raw/master/networks/models/resnet.h5), [tiny-yolo-voc](https://github.com/hollance/YOLO-CoreML-MPSNNGraph/raw/master/Convert/yad2k/model_data/tiny-yolo-voc.h5)

**CoreML Models**: [MobileNet](https://docs-assets.developer.apple.com/coreml/models/MobileNet.mlmodel), [Places205-GoogLeNet](https://docs-assets.developer.apple.com/coreml/models/GoogLeNetPlaces.mlmodel), [Inception v3](https://docs-assets.developer.apple.com/coreml/models/Inceptionv3.mlmodel)

**TensorFlow Lite Models**: [Smart Reply 1.0 ](https://storage.googleapis.com/download.tensorflow.org/models/tflite/smartreply_1.0_2017_11_01.zip), [Inception v3 2016](https://storage.googleapis.com/download.tensorflow.org/models/tflite/inception_v3_slim_2016_android_2017_11_10.zip)

**Caffe Models**: [BVLC AlexNet](http://dl.caffe.berkeleyvision.org/bvlc_alexnet.caffemodel), [BVLC CaffeNet](http://dl.caffe.berkeleyvision.org/bvlc_reference_caffenet.caffemodel), [BVLC GoogleNet](http://dl.caffe.berkeleyvision.org/bvlc_googlenet.caffemodel)

**Caffe2 Models**: [BVLC GoogleNet](https://github.com/caffe2/models/raw/master/bvlc_googlenet/predict_net.pb), [Inception v2](https://github.com/caffe2/models/raw/master/inception_v2/predict_net.pb)

**MXNet Models**: [CaffeNet](http://data.dmlc.ml/models/imagenet/squeezenet/squeezenet_v1.1-symbol.json), [SqueezeNet v1.1](https://mxnet.incubator.apache.org/model_zoo/index.html)

**TensorFlow models**: [Inception v3](https://storage.googleapis.com/download.tensorflow.org/models/inception_v3_2016_08_28_frozen.pb.tar.gz), [Inception v4](https://storage.googleapis.com/download.tensorflow.org/models/inception_v4_2016_09_09_frozen.pb.tar.gz), [Inception 5h](https://storage.googleapis.com/download.tensorflow.org/models/inception5h.zip)

## 3. 网络可视化

在客户端或者网页版中上传对应的模型结构文件，即可输出可视化结果.

例如，yolov3-tiny Darknet 模型结构可视化：

![image](https://www.aiuai.cn/uploads/sina/5ce8e0b1f228c.jpg)

Github 项目 - Netron
作者：lutzroeder
官网：https://www.lutzroeder.com/ai/

Netron 是一个强大的神经网络、深度学习和机器学习模型可视化工具. 其支持：

ONNX (.onnx, .pb, .pbtxt)
Keras (.h5, .keras)
CoreML (.mlmodel)
Caffe2 (predict_net.pb, predict_net.pbtxt)
MXNet (.model, -symbol.json)
TensorFlow Lite (.tflite)

Netron 已经测试支持的深度学习框架有：

Caffe (.caffemodel, .prototxt)
PyTorch (.pth)
Torch (.t7)
CNTK (.model, .cntk)
PaddlePaddle(__model__)
Darknet (.cfg)
scikit-learn (.pkl)
TensorFlow.js (model.json, .pb)
TensorFlow (.pb, .meta, .pbtxt)

可视化效果真心赞！

1. Netron 安装

macOS: Download the .dmg file or run brew cask install netron

Linux: Download the .AppImage or .deb file.

Windows: Download the .exe installer.

Browser: Start the browser version.

Python Server: Run pip install netron and netron -b [MODEL_FILE]. In Python run import netron and netron.start('model.onnx').

2. 下载开源模型

ONNX Models: Inception v1, Inception v2, ResNet-50, SqueezeNet

Keras Models: resnet, tiny-yolo-voc

CoreML Models: MobileNet, Places205-GoogLeNet, Inception v3

TensorFlow Lite Models: Smart Reply 1.0 , Inception v3 2016

Caffe Models: BVLC AlexNet, BVLC CaffeNet, BVLC GoogleNet

Caffe2 Models: BVLC GoogleNet, Inception v2

MXNet Models: CaffeNet, SqueezeNet v1.1

TensorFlow models: Inception v3, Inception v4, Inception 5h

3. 网络可视化

在客户端或者网页版中上传对应的模型结构文件，即可输出可视化结果.

例如，yolov3-tiny Darknet 模型结构可视化：

Last modification：January 17, 2019

© Allow specification reprint

If you think my article is useful to you, please feel free to appreciate

Loading...

Leave a Comment Cancel reply
欢迎交流学习

Comment *

Name *

🎲

Email *

Site