i.Build a CNN model (named model_0), using the following model structure for image classification. All layers should use ReLU activation except the final dense layer, which should use softmax activation.

Table 1 Layer type Conv2D MaxPooling2D Conv2D MaxPooling2D Conv2D Conv2D Conv2D Conv2D Conv2D MaxPooling2D Dropout Flatten Dense Dense Output shape (None, 141, 141, 2) (None, 70, 70, 2) (None, 70, 70, 4) (None, 35, 35, 4) (None, 35, 35, 4) (None, 35, 35, 4) (None, 35, 35, 4) (None, 35, 35, 8) (None, 35, 35, 8) (None, 17, 17, 8) (None, 17, 17, 8) (None, 2312) (None, 128) (None, 6) Param # 602 0 292 0 404 260 148 296 584 0 0 0 296064 774 Notes for Question 1b (a) (b) (c) (d)