Problem $1(30$ Points$)$: Build an convolutional neural network,the function should return model with the following specifications:

First Convolutional Layer: Uses $6$ filters with a kernel size of $55,$ 'relu' activation and

First Pooling Layer: Uses max pooling with a pool size of $22$ and a stride of $2.$

Second Convolutional Layer: Uses $16$ filters with a kernel size of $55$ and 'relu' activation.

Second Pooling Layer: Uses max pooling with the same specifications as the first pooling layer.

Flattening Layer: Converts the feature map to a $1$D array.

First Dense Layer: Has $120$ units with 'relu' activation.

Second Dense Layer: Has $84$ units with 'relu' activation.

Third Dense Layer: Has $10$ units with 'softmax' activation.

The network should be compiled using Adam optimizer, categorical$\_$crossentropy loss function and accuracy as a metric.Build an convolutional neural network,the function should return model with the following specifications:

First Convolutional Layer: Uses $6$ filters with a kernel size of $5$x$5,$ 'relu' activation and $28$x$28$ pixels, grayscale as input images size.

First Pooling Layer: Uses max pooling with a pool size of $2$x$2$ and a stride of $2.$

Second Convolutional Layer: Uses $16$ filters with a kernel size of $5$x$5$ and 'relu' activation.

Second Pooling Layer: Uses max pooling with the same specifications as the first pooling layer.

Flattening Layer: Converts the feature map to a $1$D array.

First Dense Layer: Has $120$ units with 'relu' activation.

Second Dense Layer: Has $84$ units with 'relu' activation.

Third Dense Layer: Has $10$ units with 'softmax' activation.

The network should be compiled using Adam optimizer, categorical$\_$crossentropy loss function and accuracy as a metric. from warnings import filters

import keras

from keras.models import Sequential

from keras.layers import Conv$2$D$,$ MaxPooling$2$D$,$ Flatten, Dense

from tensorflow import keras

# Function $1$: Build the model

def create$\_$model$()$:

$\text{'}\text{'}\text{'}$

output: A compile CNN model with the specifications listed in the question

$\text{'}\text{'}\text{'}$

### BEGIN SOLUTION

return model