Scribd
Upload
11 views

Tensorflow 2 Quickstart

TensorFlow

Uploaded by

Pratik Gyawali
Copyright
© © All Rights Reserved
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
11 views

Tensorflow 2 Quickstart

TensorFlow

Uploaded by

Pratik Gyawali
Copyright
© © All Rights Reserved
Available Formats
Download as TXT, PDF, TXT or read online on Scribd
You are on page 1/ 2

# -*- coding: utf-8 -*-

"""TensorFlow 2 quickstart.ipynb

Automatically generated by Colaboratory.

Original file is located at


https://colab.research.google.com/drive/1vujmOkNYhFhnBZCvYEuOs-jRhJ7ZgurE
"""

import tensorflow as tf
from tensorflow import keras
import matplotlib.pyplot as plt
import numpy as np

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

x_train = x_train / 255.0


x_test = x_test / 255.0

def visualize(images, labels ,w=2, h=5, width = 6, height = 3, color='gray',


show_axis='off'):
fig, axes = plt.subplots(w, h, figsize=(width, height))
axes = axes.flatten()

for i in range(len(images)):
axes[i].imshow(images[i], cmap=color)
axes[i].axis(show_axis)
axes[i].set_title(labels[i])

plt.tight_layout()
plt.show()

print('Display sample of train images')


visualize(x_train[:10], y_train[:10])

print('Display sample of test images')


visualize(x_test[:10], y_test[:10])

def model_(input_shape = tuple([28, 28]), activations = ['relu', 'softmax']):

return keras.Sequential([
keras.layers.Flatten(input_shape=input_shape),
keras.layers.Dense(128, activation=activations[0]),
keras.layers.Dense(10, activation=activations[1])
])

def model_train(model, data=[[],[]], epochs = 10, optimizer='adam',


loss='sparse_categorical_crossentropy', metrics=['accuracy']):
model.compile(optimizer=optimizer,
loss=loss,
metrics=metrics)
model.fit(data[0], data[1], epochs=epochs)

return model

model = model_()

print('Train the model')


model = model_train(model = model, data = [x_train, y_train], epochs=10)

print('Test Evaluation')
test_loss, test_acc = model.evaluate(x_test, y_test, verbose=2)
print('Test accuracy:', test_acc)

print('Make predictions on the test dataset')


predictions = model.predict(x_test)

print('Getting the predicted labels')


predicted_labels = np.argmax(predictions, axis=1)

# Function to visualize original images along with predicted labels


def visualize_predictions(images, true_labels, predicted_labels,
w=2, h=5, width = 8, height =
3,color='gray',show_axis='off'):
fig, axes = plt.subplots(w, h, figsize=(width, height))
axes = axes.flatten()

for i in range(len(images)):
axes[i].imshow(images[i], cmap=color)
axes[i].set_title(f"Actual: {true_labels[i]}\nPrediction:
{predicted_labels[i]}")
axes[i].axis(show_axis)

plt.tight_layout()
plt.show()

print('Visualizing original images along with predicted labels')


visualize_predictions(x_test[:10], y_test[:10], predicted_labels[:10])

You might also like