Provides pretrained computer vision models including ResNet, VGG, DenseNet, Inception, MobileNet, and SE-Net variants, all trained on ImageNet. The models are compatible with both Keras and TensorFlow Keras and can be used for transfer learning or inference. It includes multiple architecture sizes and configurations for image classification tasks.