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 sizeH
- heightW
- widthC
- 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 sizeH
- heightW
- widthC
- 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 sizeC
- 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 sizeC
- 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:
Legal Information#
The original model is distributed under the
Apache License, Version 2.0.
A copy of the license is provided in <omz_dir>/models/public/licenses/APACHE-2.0-TF-TPU.txt
.