Question 1: Build LeNet (50 pts) In this part, you will build a LeNet Model. The input to the model initialization should include the sizes of two hidden layers and the sizes of the output channels of two convolutional layers. You can refer to the original LeNet paper (Gradient-based learning applied to document recognition) for more details on the model. To improve the performance, you may consider adding batch normalization (Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift) after each convolutional layer. You may put all of your modules inside the nn. Sequential function, so it will be easy to "forward" the input. We have prepared the model initialization for you. We have used a Xavier uniform initialization for linear layers and convolutional layers