In [1]:
import tensorflow as tf
import matplotlib.pyplot as plt
import numpy as np
In [2]:
mnist = tf.keras.datasets.mnist # 28x28 images of hand-written digits 0-9

(x_train, y_train), (x_test, y_test) = mnist.load_data()

x_train = tf.keras.utils.normalize(x_train, axis = 1)
x_test = tf.keras.utils.normalize(x_test, axis = 1)

model = tf.keras.models.Sequential()
model.add(tf.keras.layers.Flatten())
model.add(tf.keras.layers.Dense(128, activation = tf.nn.relu))
model.add(tf.keras.layers.Dense(128, activation = tf.nn.relu))
model.add(tf.keras.layers.Dense(10, activation = tf.nn.softmax))

model.compile(optimizer='adam',
             loss='sparse_categorical_crossentropy',
             metrics=["accuracy"])
model.fit(x_train, y_train, epochs=3)

Epoch 1/3
1875/1875 ━━━━━━━━━━━━━━━━━━━━ 2s 744us/step - accuracy: 0.8725 - loss: 0.4654
Epoch 2/3
1875/1875 ━━━━━━━━━━━━━━━━━━━━ 1s 739us/step - accuracy: 0.9661 - loss: 0.1098
Epoch 3/3
1875/1875 ━━━━━━━━━━━━━━━━━━━━ 1s 738us/step - accuracy: 0.9781 - loss: 0.0686
Out[2]:
<keras.src.callbacks.history.History at 0x2779805a790>
In [3]:
val_loss, val_acc = model.evaluate(x_test, y_test)
print(val_loss, val_acc)

313/313 ━━━━━━━━━━━━━━━━━━━━ 0s 462us/step - accuracy: 0.9671 - loss: 0.1060
0.09289506077766418 0.9718000292778015
In [4]:
predictions = model.predict([x_test])

313/313 ━━━━━━━━━━━━━━━━━━━━ 0s 660us/step
In [5]:
for i in range(6):
        print("Number predicted :", np.argmax(predictions[i]))
        plt.imshow(x_test[i])
        plt.show()

Number predicted : 7

Number predicted : 2

Number predicted : 1

Number predicted : 0

Number predicted : 4

Number predicted : 1
In [ ]: