I am pleased to update the AlexNet and its variants to adapt to TensorFlow 2.0/2.1/2.2. Since the original script of myalexnet_forward written by the AlexNet team was build on TensorFlow 1.x in 2017, it has generated many errors during the runtime in the env. of TensorFlow 2.x. Either the AlexNet team has no time to update, or it is very hard for developers to find out the AlexNet variants to adapt to TensorFlow 2.0/2.1/2.2 with the Google search. Furthermore, I split the AlexNet model in both the OOP and the pure command style and make client applications to call the AlexNet model because the AlexNet model is not separated in most of the published scenarios which incur the difficulties for developer's deep learning. Considering that AlexNet is the greatest milestone in the CNNN/DNN history (Mr. Hinton has just won the Turing Award in 2019 and the event is not far from fading away), it is important to continue its usability and accessibility to all developers. It is exciting for me to take the opportunity to share the new update.
