学习笔记(03):英特尔®OpenVINO™工具套件中级课程--(第三章)模型优化器调优

立即学习:https://edu.csdn.net/course/play/28807/427187?utm_source=blogtoedu

OpenVINO Toolkits

模型优化器

mo.py是离线工具，为了提高性能，无关精度，有许多选项。

• 将TF/Caffe/Mxnet/ONNX转化为IR格式
• Map to supported libraray calls/kernels/layers
• 多种选项（预处理），如改变RGB通道顺序，去掉均值等
• 优化网络，合并网络结构
• 改变数据类型，FP32,FP16,INT8
• 切掉部分网络
• 支持自定义层

使用帮助

usage: mo.py [options]

optional arguments:
  -h, --help            show this help message and exit
  --framework {tf,caffe,mxnet,kaldi,onnx}
                        Name of the framework used to train the input model.

Framework-agnostic parameters:
  --input_model INPUT_MODEL, -w INPUT_MODEL, -m INPUT_MODEL
                        Tensorflow*: a file with a pre-trained model (binary
                        or text .pb file after freezing). Caffe*: a model
                        proto file with model weights
  --model_name MODEL_NAME, -n MODEL_NAME
                        Model_name parameter passed to the final create_ir
                        transform. This parameter is used to name a network in
                        a generated IR and output .xml/.bin files.
  --output_dir OUTPUT_DIR, -o OUTPUT_DIR
                        Directory that stores the generated IR. By default, it
                        is the directory from where the Model Optimizer is
                        launched.
  --input_shape INPUT_SHAPE
                        Input shape(s) that should be fed to an input node(s)
                        of the model. Shape is defined as a comma-separated
                        list of integer numbers enclosed in parentheses or
                        square brackets, for example [1,3,227,227] or
                        (1,227,227,3), where the order of dimensions depends
                        on the framework input layout of the model. For
                        example, [N,C,H,W] is used for Caffe* models and
                        [N,H,W,C] for TensorFlow* models. Model Optimizer
                        performs necessary transformations to convert the
                        shape to the layout required by Inference Engine
                        (N,C,H,W). The shape should not contain undefined
                        dimensions (? or -1) and should fit the dimensions
                        defined in the input operation of the graph. If there
                        are multiple inputs in the model, --input_shape should
                        contain definition of shape for each input separated
                        by a comma, for example: [1,3,227,227],[2,4] for a
                        model with two inputs with 4D and 2D shapes.
                        Alternatively, you can specify shapes with the --input
                        option.
  --scale SCALE, -s SCALE
                        All input values coming from original network inputs
                        will be divided by this value. When a list of inputs
                        is overridden by the --input parameter, this scale is
                        not applied for any input that does not match with the
                        original input of the model.
  --reverse_input_channels
                        Switch the input channels order from RGB to BGR (or
                        vice versa). Applied to original inputs of the model
                        if and only if a number of channels equals 3. Applied
                        after application of --mean_values and --scale_values
                        options, so numbers in --mean_values and
                        --scale_values go in the order of channels used in the
                        original model.
  --log_level {CRITICAL,ERROR,WARN,WARNING,INFO,DEBUG,NOTSET}
                        Logger level
  --input INPUT         Quoted list of comma-separated input nodes names with
                        shapes, data types, and values for freezing. The shape
                        and value are specified as space-separated lists. The
                        data type of input node is specified in braces and can
                        have one of the values: f64 (float64), f32 (float32),
                        f16 (float16), i64 (int64), i32 (int32), u8 (uint8),
                        boolean. For example, use the following format to set
                        input port 0 of the node `node_name1` with the shape
                        [3 4] as an input node and freeze output port 1 of the
                        node `node_name2` with the value [20 15] of int32