Tools OV MO

[junxnone]

Model Optimizer

Brief

• MO - Model Optimizer 把各种框架训练的模型文件转换成 OpenVINO 可以识别的文件并优化模型
• Support Framework : Tensorflow/MXNet/Kaldi/ONNX
  • Pytorch ==> ONNX ==> IR

Parameters	Description
--reverse_input_channels	当输入为 3 通道图像时, 转换通道顺序 RGB -> BGR
--input_shape	输入的 Shape NCHW/NHWC
--mean_values	RGB Mean values ==> image - mean
--scale	归一化 ==> image - mean) /scale

models mean/scale value 和 Training 保持一致

UseCase

SSD_Mobilenet

• ssd_mobilenet_v2_coco
• ssd_mobilenet_v1_fpn_640

python3 /opt/intel/openvino/deployment_tools/model_optimizer/mo.py \
        --input_model frozen_inference_graph.pb \
        --tensorflow_use_custom_operations_config /opt/intel/openvino/deployment_tools/model_optimizer/extensions/front/tf/ssd_v2_support.json \
        --tensorflow_object_detection_api_pipeline_config pipeline.config \
        --data_type FP16\
        --output_dir FP16

YOLO

• YOLO v3
  • /opt/intel/openvino/deployment_tools/model_optimizer/extensions/front/tf/yolo_v3.json

    python3 /opt/intel/openvino/deployment_tools/model_optimizer/mo_tf.py -b 1 \
    --input_model ./frozen_darknet_yolov3_model.pb \
    --tensorflow_use_custom_operations_config ./yolo_v3.json \
    --data_type FP16\
    --output_dir ./FP16

• YOLO tiny
  • /opt/intel/openvino/deployment_tools/model_optimizer/extensions/front/tf/yolo_v3_tiny.json

    python3 /opt/intel/openvino/deployment_tools/model_optimizer/mo_tf.py -b 1 \
    --input_model ./frozen_darknet_yolov3_model.pb \
    --tensorflow_use_custom_operations_config ./yolo_v3_tiny.json \
    --data_type FP16\
    --output_dir ./tiny_FP16

[junxnone]

junxnone/tech-io#17 junxnone/tech-io#700