mixnet-l

Use Case and High-Level Description

MixNets are a family of mobile-sizes image classification models equipped with MixConv, a new type of mixed depthwise convolutions. There are three MixNet architectures - MixNet-S (Small), MixNet-M (Middle), MixNet-L (Large). The main differences are using MixConv with different kernel sizes and number of layers. Using MixNet-L allows to achieve greater accuracy. All the MixNet models have been pretrained on the ImageNet dataset. For details about this family of models, check out the TensorFlow Cloud TPU repository and paper.

Specification

Metric

Value

Type

Classification

GFLOPs

0.565

MParams

7.300

Source framework

TensorFlow*

Accuracy

Metric

Original model

Converted model

Top 1

78.30%

78.30%

Top 5

93.91%

93.91%

Input

Original Model

Image, name - image, shape - 1, 224, 224, 3, format is B, H, W, C, where:

  • B - batch size

  • H - height

  • W - width

  • C - channel

Channel order is RGB.

Converted Model

Image, name - IteratorGetNext/placeholder_out_port_0, shape - 1, 224, 224, 3, format is B, H, W, C, where:

  • B - batch size

  • H - height

  • W - width

  • C - channel

Channel order is BGR.

Output

Original Model

Object classifier according to ImageNet classes, name - logits, shape - 1,1000, output data format is B,C where:

  • B - batch size

  • C - predicted logits for each class

Converted Model

Object classifier according to ImageNet classes, name - logits, shape - 1,1000, output data format is B,C where:

  • B - batch size

  • C - predicted logits for each class

Download a Model and Convert it into OpenVINO™ IR Format

You can download models and if necessary convert them into OpenVINO™ IR format using the Model Downloader and other automation tools as shown in the examples below.

An example of using the Model Downloader:

omz_downloader --name <model_name>

An example of using the Model Converter:

omz_converter --name <model_name>

Demo usage

The model can be used in the following demos provided by the Open Model Zoo to show its capabilities: