Convert Full ImageNet Pre-trained Model from MXNet to PyTorch

Currently there are many available deep learning frameworks for researchers and engineers to implement their desired deep models. Each deep learning framework has its own advantages and disadvantages. For example, TensorFlow has a great community, PyTorch is an excellent framework to easily develop models in a short time and also it provides a fantastic C++ API for production level tasks, MXNet is a great framework for extremely large-scale training (i.e., it is an ultra-scalable framework which can speedup training time on distributed systems and multiple GPUs), and so on.

This is a companion discussion topic for the original entry at