Mobilenetv2

A model that embedded with Multi-Task-Learning-training

Implementation

• Replace every Conv2d and BN2d Layer in pytorch with layer_node.Conv2dNode, or layer_node.BN2dNode, with your interested task and wrap Sequential using layer_node.Sequential
• Inherited from mtl_model.mtl_model and call self.compute_depth() at the end of __init__().