Learning curves

Introduction to Deep Learning with Keras

Miguel Esteban

Data Scientist & Founder

Introduction to Deep Learning with Keras

Introduction to Deep Learning with Keras

Introduction to Deep Learning with Keras

Introduction to Deep Learning with Keras

Introduction to Deep Learning with Keras

Introduction to Deep Learning with Keras

Introduction to Deep Learning with Keras

Introduction to Deep Learning with Keras
# Store initial model weights
init_weights = model.get_weights()

# Lists for storing accuracies
train_accs = []
tests_accs = []
Introduction to Deep Learning with Keras
for train_size in train_sizes:
    # Split a fraction according to train_size
    X_train_frac, _, y_train_frac, _ = 
    train_test_split(X_train, y_train, train_size=train_size)

    # Set model initial weights
    model.set_weights(initial_weights)

    # Fit model on the training set fraction
    model.fit(X_train_frac, y_train_frac, epochs=100,
                  verbose=0,
                  callbacks=[EarlyStopping(monitor='loss', patience=1)])

    # Get the accuracy for this training set fraction
    train_acc = model.evaluate(X_train_frac, y_train_frac, verbose=0)[1]
    train_accs.append(train_acc)

    # Get the accuracy on the whole test set
    test_acc = model.evaluate(X_test, y_test, verbose=0)[1]
    test_accs.append(test_acc)
    print("Done with size: ", train_size)
Introduction to Deep Learning with Keras

Time to dominate all curves!

Introduction to Deep Learning with Keras

Preparing Video For Download...