TensorFlow pretrained ImageNet models for Foolbox

This repository provides the Keras Applications (pretrained ImageNet models) in a Foolbox Native compatible format.

This code requires Foolbox 3.0 or newer.

Example

import foolbox as fbn

url = "https://github.com/jonasrauber/foolbox-tensorflow-keras-applications"
fmodel = fbn.zoo.get_model(url, name="MobileNetV2")

images, labels = fbn.samples(fmodel, dataset='imagenet', batchsize=16)
print(fbn.accuracy(fmodel, images, labels))
# -> 0.9375

# you can now attack fmodel using Foolbox attacks
# ...

Supported Models

• DenseNet121
• DenseNet169
• DenseNet201
• InceptionResNetV2
• InceptionV3
• MobileNet
• MobileNetV2
• NASNetLarge
• NASNetMobile
• ResNet50
• ResNet101
• ResNet152
• ResNet50V2
• ResNet101V2
• ResNet152V2
• VGG16
• VGG19
• Xception