densenet-169¶

Use Case and High-Level Description¶

The densenet-169 model is one of the DenseNet group of models designed to perform image classification. The main difference with the densenet-121 model is the size and accuracy of the model. The densenet-169 is larger at just about 55MB in size vs the densenet-121 model’s roughly 31MB size. Originally trained on Torch, the authors converted them into Caffe* format. All the DenseNet models have been pre-trained on the ImageNet image database. For details about this family of models, check out the repository.

The model input is a blob that consists of a single image of 1, 3, 224, 224 in BGR order. The BGR mean values need to be subtracted as follows: [103.94, 116.78, 123.68] before passing the image blob into the network. In addition, values must be divided by 0.017.

The model output for densenet-169 is the typical object classifier output for the 1000 different classifications matching those in the ImageNet database.

Metric

Value

Type

Classification

GFLOPs

6.788

MParams

14.139

Source framework

Caffe*

Metric

Value

Top 1

76.106%

Top 5

93.106%

Input¶

Original model¶

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

• B - batch size

• C - channel

• H - height

• W - width

Channel order is BGR. Mean values - [103.94, 116.78, 123.68], scale value - 58.8235294117647.

Converted model¶

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

• B - batch size

• C - channel

• H - height

• W - width

Channel order is BGR.

Output¶

Original model¶

Object classifier according to ImageNet classes, name - fc6, shape - 1, 1000, 1, 1, contains predicted probability for each class in logits format.

Converted model¶

Object classifier according to ImageNet classes, name - fc6, shape - 1, 1000, 1, 1, contains predicted probability for each class in logits format.

python3 <omz_dir>/tools/downloader/downloader.py --name <model_name>
python3 <omz_dir>/tools/downloader/converter.py --name <model_name>