目录

Open Model Zoo automation tools

omz_downloader

用于访问预训练深度学习公共模型和英特尔训练模型的集合。

omz_converter

使用模型优化器将存储在原始深度学习框架格式中的 Open Model Zoo 模型转换为 OpenVINO 中间表示 (IR)。

omz_quantizer

用于使用训练后优化工具（Post-Training Optimization Tool）将 IR 格式的全精度模型自动量化为低精度版本。

omz_info_dumper

用于将有关模型的信息转储为机器可读格式。

omz_data_downloader

数据集下载器

查看可用的模型

Open Model Zoo 自动化工具都可以使用参数 --print_all

$ omz_tool --print_all

Sphereface
aclnet
aclnet-int8
action-recognition-0001
age-gender-recognition-retail-0013
alexnet
......
yolo-v3-onnx
yolo-v3-tf
yolo-v3-tiny-onnx
yolo-v3-tiny-tf
yolo-v4-tf
yolo-v4-tiny-tf
yolof
yolox-tiny

Demos

Open Model Zoo Demos

Clone Open Model Zoo

git clone https://github.com/openvinotoolkit/open_model_zoo.git
cd open_model_zoo
git submodule update --init --recursive

Python Demos

安装依赖包

Installing Dependencies for Python Demos
pip install -r demos/requirements.txt -i https://mirrors.aliyun.com/pypi/simple/
Installing Python Model API package
pip install demos/common/python

Face Detection MTCNN Python* Demo

cd demos/face_detection_mtcnn_demo/python
模型下载
omz_downloader --list models.lst
模型转换
omz_converter --list models.lst
下载图片
wget https://manastirea.petru-voda.ro/wp-content/uploads/2017/09/tineri.jpg -O human-crowd.jpg
wget http://ptvnews.ph/wp-content/uploads/2018/12/img7744-1920x1266.jpg -O people.jpg
wget https://www.lse.ac.uk/government/Assets/Images/People/Department-photo/people-page-thumbnail-1200x600.jpg -O people1.jpg
人脸检测推理
  • Windows 
    python face_detection_mtcnn_demo.py `
  -i people.jpg `
  -m_p public/mtcnn/mtcnn-p/FP32/mtcnn-p.xml `
  -m_o public/mtcnn/mtcnn-o/FP32/mtcnn-o.xml `
  -m_r public/mtcnn/mtcnn-r/FP32/mtcnn-r.xml `
  -th 0.7
  • Linux 
    python face_detection_mtcnn_demo.py \
  -i people.jpg \
  -m_p public/mtcnn/mtcnn-p/FP32/mtcnn-p.xml \
  -m_o public/mtcnn/mtcnn-o/FP32/mtcnn-o.xml \
  -m_r public/mtcnn/mtcnn-r/FP32/mtcnn-r.xml \
  -th 0.7

[ INFO ] OpenVINO Runtime
[ INFO ] 	build: 2022.1.0-7019-cdb9bec7210-releases/2022/1
[ INFO ] Reading Proposal model public/mtcnn/mtcnn-p/FP32/mtcnn-p.xml
[ INFO ] Reading Refine model public/mtcnn/mtcnn-r/FP32/mtcnn-r.xml
[ INFO ] Reading Output model public/mtcnn/mtcnn-o/FP32/mtcnn-o.xml
[ INFO ] The Proposal model public/mtcnn/mtcnn-p/FP32/mtcnn-p.xml is loaded to CPU
[ INFO ] The Refine model public/mtcnn/mtcnn-r/FP32/mtcnn-r.xml is loaded to CPU
[ INFO ] The Output model public/mtcnn/mtcnn-o/FP32/mtcnn-o.xml is loaded to CPU
[ INFO ] Metrics report:
[ INFO ] 	Latency: 1160.0 ms
[ INFO ] 	FPS: 0.9

如果在服务器上运行，可以使用参数 --no_show；需要保存输出结果，可以使用参数 -o filename.jpg。

参考资料