1. Consider the following deep convolutional neural network: Convolutional input layer, 32 feature maps with a size of 3 x 3 and a rectifier activation function. Dropout laver at 20%. Convolutional layer, 32 feature maps with a size of 3 × 3 and a rectifier activation function. Max Pool layer with size 2 × 2. Convolutional laver, 64 feature maps with a size of 3 × 3 and a rectifier activation function. Dropout layer at 20%. Convolutional layer, 64 feature maps with a size of 3 × 3 and a rectifier activation function. Max Pool laver with size 2 × 2. Convolutional layer, 128 feature maps with a size of 3 × 3 and a rectifier activation function. Dropout layer at 20%. Convolutional layer, 128 feature maps with a size of 3 × 3 and a rectifier activation function. Max Pool laver with size 2 × 2. Flatten layer. Dropout layer at 20%. Fully connected layer with 1,024 units and a rectifier activation function. Dropout laver at 20%. Fully connected layer with 512 units and a rectifier activation function. Dropout layer at 20%. Fully connected output layer with 10 units and a softmax activation function. Use this network to perform a n-class classification job on the CIFAR 10 dataset. CIFAR10: from keras.datasets import cifar10 Other specifications: # fix random seed for reproducibility seed = 7 numpy.random. seed (seed) # load data (X_train, y-train), (X_test, y-test) = cifar10.load_data () Report the training and testing accuracies. What are some possible ways to improve the performance of your model?

1. Consider the following deep convolutional neural network:

1. Convolutional input layer, 32 feature maps with a size of 3 x 3 and a rectifier activation function.
2. Dropout laver at 20%.
3. Convolutional layer, 32 feature maps with a size of 3 × 3 and a rectifier activation function.
4. Max Pool layer with size 2 × 2.
5. Convolutional laver, 64 feature maps with a size of 3 × 3 and a rectifier activation function.
6. Dropout layer at 20%.
7. Convolutional layer, 64 feature maps with a size of 3 × 3 and a rectifier activation function.
8. Max Pool laver with size 2 × 2.
9. Convolutional layer, 128 feature maps with a size of 3 × 3 and a rectifier activation function.
10. Dropout layer at 20%.
11. Convolutional layer, 128 feature maps with a size of 3 × 3 and a rectifier activation function.
12. Max Pool laver with size 2 × 2.
13. Flatten layer.
14. Dropout layer at 20%.
15. Fully connected layer with 1,024 units and a rectifier activation function.
16. Dropout laver at 20%.
17. Fully connected layer with 512 units and a rectifier activation function.
18. Dropout layer at 20%.
19. Fully connected output layer with 10 units and a softmax activation function.

Use this network to perform a n-class classification job on the CIFAR 10 dataset.

CIFAR10:

from keras.datasets import cifar10

Other specifications:

# fix random seed for reproducibility

seed = 7

numpy.random. seed (seed)

# load data

(X_train, y-train), (X_test, y-test) = cifar10.load_data ()

Report the training and testing accuracies. What are some possible ways to improve the performance of your model?