The EfficientNet effectively conducts the model scaling and balance network depth, width, and resolution for the better performance. The reserchers uniformly scales all dimensions by using a simple but effective compound coefficient.
Based on the effectiveness of both MobileNets and ResNet, NasNet(neural architecture search) is uded to design a new baseline network and scale it up to obtain a family of models, called EfficientNets B0-B7, which achieve better accuracy and efficiency.
Make the necessary changes to adapt to the environment of TensorFlow 2.3, Keras 2.4.3, CUDA Toolkit 11.0, cuDNN 8.0.1 and CUDA 450.57. In addition, write the new lines of code to replace the deprecated code. I would like to thank all of the creators and interptretors for their contributions.