The code in this repo enables you to download the benchmark deep learning Keras models as files saved to your system. All these benchmark models are pretrained on the imagenet dataset.
Each available network is identified by a unique I.D. , the type of network can be infereded from the i.d.

Thanks to Adrian Rosebrock for his inspiring post https://www.pyimagesearch.com/2017/03/20/imagenet-vggnet-resnet-inception-xception-keras/

These are the list of avilabile I.D.s: ["InceptionV3","ResNet50","VGG16","VGG19","NASNetMobile","NASNetLarge","InceptionResNetV2","MobileNet","DenseNet121","DenseNet169","DenseNet201"]

You can choose a certain model to save to your computer (for ex InceptionV3)
python deepModelsSaver.py --networkName InceptionV3

you can also download all the models to your computer as follows
python deepModelsSaver.py --networkName all

you can choose to classify a certain image with an imagenet label with a certain model as follows:
python benchMarkModelPredictor.py --networkName NASNetLarge --img_path images/dog_beagle.png

you can choose to classify a collection of images in a folder with an imagenet label with a certain model as follows:

python benchMarkModelPredictorFolder.py --networkName NASNetLarge --folder_path imagenet_images

A video will be created showing results of classification.

Imagenet images were downloaded using https://github.com/mf1024/ImageNet-Datasets-Downloader