10/16/24, 2:59 PM DL2 - Jupyter Notebook

Implementing Feedforward neural networks with Keras and TensorFlow a. Import the
necessary packages b. Load the training and testing data (MNIST/CIFAR10) c. Define the
network architecture using Keras d. Train the model using SGD e. Evaluate the network f.
Plot the training loss and accuracy

In [2]: #installations
from sklearn.preprocessing import LabelBinarizer
from sklearn.metrics import classification_report
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import Dense
from tensorflow.keras.optimizers import SGD
from tensorflow.keras.datasets import mnist
from tensorflow.keras import backend as K
import matplotlib.pyplot as plt
import numpy as np
import warnings
warnings.filterwarnings("ignore")

In [3]: #grabbing the mnist dataset

((X_train, Y_train), (X_test, Y_test)) = mnist.load_data()
X_train = X_train.reshape((X_train.shape[0], 28 * 28 * 1))
X_test = X_test.reshape((X_test.shape[0], 28 * 28 * 1))
X_train = X_train.astype("float32") / 255.0
X_test = X_test.astype("float32") / 255.0

In [4]: lb = LabelBinarizer()
Y_train = lb.fit_transform(Y_train)
Y_test = lb.transform(Y_test)

In [9]: #building the model

model = Sequential()
model.add(Dense(128, input_shape=(784,), activation="sigmoid"))
model.add(Dense(64, activation="sigmoid"))
model.add(Dense(10, activation="softmax"))
model.summary()

Model: "sequential_2"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense_6 (Dense) (None, 128) 100480

dense_7 (Dense) (None, 64) 8256

dense_8 (Dense) (None, 10) 650

=================================================================
Total params: 109386 (427.29 KB)
Trainable params: 109386 (427.29 KB)
Non-trainable params: 0 (0.00 Byte)
_________________________________________________________________

localhost:8888/notebooks/Downloads/Deep-learning-main/DL2.ipynb 1/5
10/16/24, 2:59 PM DL2 - Jupyter Notebook

In [10]: sgd = SGD(0.02)

epochs=20
model.compile(loss="categorical_crossentropy", optimizer=sgd,metrics=["accu
H = model.fit(X_train, Y_train, validation_data=(X_test, Y_test),epochs=epo

localhost:8888/notebooks/Downloads/Deep-learning-main/DL2.ipynb 2/5
10/16/24, 2:59 PM DL2 - Jupyter Notebook

Epoch 1/20
469/469 [==============================] - 2s 3ms/step - loss: 2.2552 - ac
curacy: 0.2486 - val_loss: 2.1963 - val_accuracy: 0.4051
Epoch 2/20
469/469 [==============================] - 1s 2ms/step - loss: 2.1205 - ac
curacy: 0.4956 - val_loss: 2.0132 - val_accuracy: 0.5698
Epoch 3/20
469/469 [==============================] - 1s 2ms/step - loss: 1.8644 - ac
curacy: 0.5861 - val_loss: 1.6742 - val_accuracy: 0.6286
Epoch 4/20
469/469 [==============================] - 1s 2ms/step - loss: 1.4991 - ac
curacy: 0.6398 - val_loss: 1.3131 - val_accuracy: 0.6951
Epoch 5/20
469/469 [==============================] - 1s 2ms/step - loss: 1.1908 - ac
curacy: 0.7068 - val_loss: 1.0600 - val_accuracy: 0.7372
Epoch 6/20
469/469 [==============================] - 1s 2ms/step - loss: 0.9822 - ac
curacy: 0.7592 - val_loss: 0.8896 - val_accuracy: 0.7862
Epoch 7/20
469/469 [==============================] - 1s 2ms/step - loss: 0.8384 - ac
curacy: 0.7952 - val_loss: 0.7692 - val_accuracy: 0.8129
Epoch 8/20
469/469 [==============================] - 1s 2ms/step - loss: 0.7343 - ac
curacy: 0.8203 - val_loss: 0.6796 - val_accuracy: 0.8298
Epoch 9/20
469/469 [==============================] - 1s 2ms/step - loss: 0.6570 - ac
curacy: 0.8366 - val_loss: 0.6134 - val_accuracy: 0.8437
Epoch 10/20
469/469 [==============================] - 1s 3ms/step - loss: 0.5985 - ac
curacy: 0.8493 - val_loss: 0.5619 - val_accuracy: 0.8542
Epoch 11/20
469/469 [==============================] - 1s 2ms/step - loss: 0.5533 - ac
curacy: 0.8577 - val_loss: 0.5213 - val_accuracy: 0.8630
Epoch 12/20
469/469 [==============================] - 1s 2ms/step - loss: 0.5174 - ac
curacy: 0.8657 - val_loss: 0.4897 - val_accuracy: 0.8719
Epoch 13/20
469/469 [==============================] - 1s 3ms/step - loss: 0.4881 - ac
curacy: 0.8717 - val_loss: 0.4630 - val_accuracy: 0.8773
Epoch 14/20
469/469 [==============================] - 1s 2ms/step - loss: 0.4642 - ac
curacy: 0.8769 - val_loss: 0.4419 - val_accuracy: 0.8803
Epoch 15/20
469/469 [==============================] - 1s 2ms/step - loss: 0.4442 - ac
curacy: 0.8818 - val_loss: 0.4233 - val_accuracy: 0.8865
Epoch 16/20
469/469 [==============================] - 1s 3ms/step - loss: 0.4274 - ac
curacy: 0.8856 - val_loss: 0.4074 - val_accuracy: 0.8891
Epoch 17/20
469/469 [==============================] - 1s 2ms/step - loss: 0.4131 - ac
curacy: 0.8887 - val_loss: 0.3945 - val_accuracy: 0.8930
Epoch 18/20
469/469 [==============================] - 1s 2ms/step - loss: 0.4006 - ac
curacy: 0.8914 - val_loss: 0.3838 - val_accuracy: 0.8957
Epoch 19/20
469/469 [==============================] - 1s 3ms/step - loss: 0.3900 - ac
curacy: 0.8932 - val_loss: 0.3738 - val_accuracy: 0.8967
Epoch 20/20
469/469 [==============================] - 1s 2ms/step - loss: 0.3805 - ac
curacy: 0.8961 - val_loss: 0.3646 - val_accuracy: 0.8982

localhost:8888/notebooks/Downloads/Deep-learning-main/DL2.ipynb 3/5
10/16/24, 2:59 PM DL2 - Jupyter Notebook

In [11]: #making the predictions

predictions = model.predict(X_test, batch_size=128)
print(classification_report(Y_test.argmax(axis=1),predictions.argmax(axis=1

79/79 [==============================] - 0s 2ms/step

precision recall f1-score support

0 0.92 0.97 0.95 980

1 0.96 0.97 0.96 1135
2 0.89 0.88 0.88 1032
3 0.89 0.87 0.88 1010
4 0.88 0.92 0.90 982
5 0.86 0.81 0.84 892
6 0.90 0.93 0.92 958
7 0.91 0.90 0.91 1028
8 0.86 0.85 0.86 974
9 0.89 0.86 0.87 1009

accuracy 0.90 10000

macro avg 0.90 0.90 0.90 10000
weighted avg 0.90 0.90 0.90 10000

In [12]: #plotting the training loss and accuracy

plt.style.use("ggplot")
plt.figure()
plt.plot(np.arange(0, epochs), H.history["loss"], label="train_loss")
plt.plot(np.arange(0, epochs), H.history["val_loss"], label="val_loss")
plt.plot(np.arange(0, epochs), H.history["accuracy"], label="train_acc")
plt.plot(np.arange(0, epochs), H.history["val_accuracy"], label="val_acc")
plt.title("Training Loss and Accuracy")
plt.xlabel("Epoch #")
plt.ylabel("Loss/Accuracy")
plt.legend()

Out[12]: <matplotlib.legend.Legend at 0x256b7f48850>

localhost:8888/notebooks/Downloads/Deep-learning-main/DL2.ipynb 4/5
10/16/24, 2:59 PM DL2 - Jupyter Notebook

localhost:8888/notebooks/Downloads/Deep-learning-main/DL2.ipynb 5/5