def model_builder(n_inputs, n_outputs):
model = NeuralNetwork(optimizer=Adam(), loss=CrossEntropy)
model.add(Dense(16, input_shape=(n_inputs, )))
model.add(Activation('relu'))
model.add(Dense(n_outputs))
model.add(Activation('softmax'))
return model
def model_builder(n_inputs, n_outputs):
model = NeuralNetwork(optimizer=Adam(), loss=CrossEntropy)
model.add(Dense(16, input_shape=(n_inputs,)))
model.add(Activation('relu'))
model.add(Dense(n_outputs))
model.add(Activation('softmax'))
return model
def __init__(self):
self.img_rows = 28
self.img_cols = 28
self.img_dim = self.img_rows * self.img_cols
self.latent_dim = 128 # The dimension of the data embedding
optimizer = Adam(learning_rate=0.0002, b1=0.5)
loss_function = SquareLoss
self.encoder = self.build_encoder(optimizer, loss_function)
self.decoder = self.build_decoder(optimizer, loss_function)
self.autoencoder = NeuralNetwork(optimizer=optimizer, loss=loss_function)
self.autoencoder.layers.extend(self.encoder.layers)
self.autoencoder.layers.extend(self.decoder.layers)
print ()
self.autoencoder.summary(name="Variational Autoencoder")
def __init__(self):
self.img_rows = 28
self.img_cols = 28
self.img_dim = self.img_rows * self.img_cols
self.latent_dim = 100
optimizer = Adam(learning_rate=0.0002, b1=0.5)
loss_function = CrossEntropy
# Build the discriminator
self.discriminator = self.build_discriminator(optimizer, loss_function)
# Build the generator
self.generator = self.build_generator(optimizer, loss_function)
# Build the combined model
self.combined = NeuralNetwork(optimizer=optimizer, loss=loss_function)
self.combined.layers.extend(self.generator.layers)
self.combined.layers.extend(self.discriminator.layers)
print ()
self.generator.summary(name="Generator")
self.discriminator.summary(name="Discriminator")
def build_discriminator(self, optimizer, loss_function):
model = NeuralNetwork(optimizer=optimizer, loss=loss_function)
model.add(
Conv2D(32,
filter_shape=(3, 3),
stride=2,
input_shape=self.img_shape,
padding='same'))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.25))
model.add(Conv2D(64, filter_shape=(3, 3), stride=2, padding='same'))
model.add(ZeroPadding2D(padding=((0, 1), (0, 1))))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.25))
model.add(BatchNormalization(momentum=0.8))
model.add(Conv2D(128, filter_shape=(3, 3), stride=2, padding='same'))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.25))
model.add(BatchNormalization(momentum=0.8))
model.add(Conv2D(256, filter_shape=(3, 3), stride=1, padding='same'))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(2))
model.add(Activation('softmax'))
return model
def build_generator(self, optimizer, loss_function):
model = NeuralNetwork(optimizer=optimizer, loss=loss_function)
model.add(Dense(128 * 7 * 7, input_shape=(100, )))
model.add(Activation('leaky_relu'))
model.add(Reshape((128, 7, 7)))
model.add(BatchNormalization(momentum=0.8))
model.add(UpSampling2D())
model.add(Conv2D(128, filter_shape=(3, 3), padding='same'))
model.add(Activation("leaky_relu"))
model.add(BatchNormalization(momentum=0.8))
model.add(UpSampling2D())
model.add(Conv2D(64, filter_shape=(3, 3), padding='same'))
model.add(Activation("leaky_relu"))
model.add(BatchNormalization(momentum=0.8))
model.add(Conv2D(1, filter_shape=(3, 3), padding='same'))
model.add(Activation("tanh"))
return model
class DCGAN():
def __init__(self):
self.img_rows = 28
self.img_cols = 28
self.channels = 1
self.img_shape = (self.channels, self.img_rows, self.img_cols)
self.latent_dim = 100
optimizer = Adam(learning_rate=0.0002, b1=0.5)
loss_function = CrossEntropy
# Build the discriminator
self.discriminator = self.build_discriminator(optimizer, loss_function)
# Build the generator
self.generator = self.build_generator(optimizer, loss_function)
# Build the combined model
self.combined = NeuralNetwork(optimizer=optimizer, loss=loss_function)
self.combined.layers.extend(self.generator.layers)
self.combined.layers.extend(self.discriminator.layers)
print()
self.generator.summary(name="Generator")
self.discriminator.summary(name="Discriminator")
def build_generator(self, optimizer, loss_function):
model = NeuralNetwork(optimizer=optimizer, loss=loss_function)
model.add(Dense(128 * 7 * 7, input_shape=(100, )))
model.add(Activation('leaky_relu'))
model.add(Reshape((128, 7, 7)))
model.add(BatchNormalization(momentum=0.8))
model.add(UpSampling2D())
model.add(Conv2D(128, filter_shape=(3, 3), padding='same'))
model.add(Activation("leaky_relu"))
model.add(BatchNormalization(momentum=0.8))
model.add(UpSampling2D())
model.add(Conv2D(64, filter_shape=(3, 3), padding='same'))
model.add(Activation("leaky_relu"))
model.add(BatchNormalization(momentum=0.8))
model.add(Conv2D(1, filter_shape=(3, 3), padding='same'))
model.add(Activation("tanh"))
return model
def build_discriminator(self, optimizer, loss_function):
model = NeuralNetwork(optimizer=optimizer, loss=loss_function)
model.add(
Conv2D(32,
filter_shape=(3, 3),
stride=2,
input_shape=self.img_shape,
padding='same'))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.25))
model.add(Conv2D(64, filter_shape=(3, 3), stride=2, padding='same'))
model.add(ZeroPadding2D(padding=((0, 1), (0, 1))))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.25))
model.add(BatchNormalization(momentum=0.8))
model.add(Conv2D(128, filter_shape=(3, 3), stride=2, padding='same'))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.25))
model.add(BatchNormalization(momentum=0.8))
model.add(Conv2D(256, filter_shape=(3, 3), stride=1, padding='same'))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.25))
model.add(Flatten())
model.add(Dense(2))
model.add(Activation('softmax'))
return model
def train(self, epochs, batch_size=128, save_interval=50):
mnist = fetch_mldata('MNIST original')
X = mnist.data.reshape((-1, ) + self.img_shape)
y = mnist.target
# Rescale -1 to 1
X = (X.astype(np.float32) - 127.5) / 127.5
half_batch = int(batch_size / 2)
for epoch in range(epochs):
# ---------------------
# Train Discriminator
# ---------------------
self.discriminator.set_trainable(True)
# Select a random half batch of images
idx = np.random.randint(0, X.shape[0], half_batch)
imgs = X[idx]
# Sample noise to use as generator input
noise = np.random.normal(0, 1, (half_batch, 100))
# Generate a half batch of images
gen_imgs = self.generator.predict(noise)
valid = np.concatenate((np.ones(
(half_batch, 1)), np.zeros((half_batch, 1))),
axis=1)
fake = np.concatenate((np.zeros(
(half_batch, 1)), np.ones((half_batch, 1))),
axis=1)
# Train the discriminator
d_loss_real, d_acc_real = self.discriminator.train_on_batch(
imgs, valid)
d_loss_fake, d_acc_fake = self.discriminator.train_on_batch(
gen_imgs, fake)
d_loss = 0.5 * (d_loss_real + d_loss_fake)
d_acc = 0.5 * (d_acc_real + d_acc_fake)
# ---------------------
# Train Generator
# ---------------------
# We only want to train the generator for the combined model
self.discriminator.set_trainable(False)
# Sample noise and use as generator input
noise = np.random.normal(0, 1, (batch_size, self.latent_dim))
# The generator wants the discriminator to label the generated samples as valid
valid = np.concatenate((np.ones(
(batch_size, 1)), np.zeros((batch_size, 1))),
axis=1)
# Train the generator
g_loss, g_acc = self.combined.train_on_batch(noise, valid)
# Display the progress
print("%d [D loss: %f, acc: %.2f%%] [G loss: %f, acc: %.2f%%]" %
(epoch, d_loss, 100 * d_acc, g_loss, 100 * g_acc))
# If at save interval => save generated image samples
if epoch % save_interval == 0:
self.save_imgs(epoch)
def save_imgs(self, epoch):
r, c = 5, 5
noise = np.random.normal(0, 1, (r * c, 100))
gen_imgs = self.generator.predict(noise)
# Rescale images 0 - 1 (from -1 to 1)
gen_imgs = 0.5 * (gen_imgs + 1)
fig, axs = plt.subplots(r, c)
plt.suptitle("Deep Convolutional Generative Adversarial Network")
cnt = 0
for i in range(r):
for j in range(c):
axs[i, j].imshow(gen_imgs[cnt, 0, :, :], cmap='gray')
axs[i, j].axis('off')
cnt += 1
fig.savefig("mnist_%d.png" % epoch)
plt.close()
def build_decoder(self, optimizer, loss_function):
decoder = NeuralNetwork(optimizer=optimizer, loss=loss_function)
decoder.add(Dense(256, input_shape=(self.latent_dim, )))
decoder.add(Activation('leaky_relu'))
decoder.add(BatchNormalization(momentum=0.8))
decoder.add(Dense(512))
decoder.add(Activation('leaky_relu'))
decoder.add(BatchNormalization(momentum=0.8))
decoder.add(Dense(self.img_dim))
decoder.add(Activation('tanh'))
return decoder
class Autoencoder(object):
"""An Autoencoder with deep fully-connected neural nets.
Training Data: MNIST Handwritten Digits (28x28 images)
"""
def __init__(self):
self.img_rows = 28
self.img_cols = 28
self.img_dim = self.img_rows * self.img_cols
self.latent_dim = 128 # The dimension of the data embedding
optimizer = Adam(learning_rate=0.0002, b1=0.5)
loss_function = SquareLoss
self.encoder = self.build_encoder(optimizer, loss_function)
self.decoder = self.build_decoder(optimizer, loss_function)
self.autoencoder = NeuralNetwork(optimizer=optimizer,
loss=loss_function)
self.autoencoder.layers.extend(self.encoder.layers)
self.autoencoder.layers.extend(self.decoder.layers)
print()
self.autoencoder.summary(name="Variational Autoencoder")
def build_encoder(self, optimizer, loss_function):
encoder = NeuralNetwork(optimizer=optimizer, loss=loss_function)
encoder.add(Dense(512, input_shape=(self.img_dim, )))
encoder.add(Activation('leaky_relu'))
encoder.add(BatchNormalization(momentum=0.8))
encoder.add(Dense(256))
encoder.add(Activation('leaky_relu'))
encoder.add(BatchNormalization(momentum=0.8))
encoder.add(Dense(self.latent_dim))
return encoder
def build_decoder(self, optimizer, loss_function):
decoder = NeuralNetwork(optimizer=optimizer, loss=loss_function)
decoder.add(Dense(256, input_shape=(self.latent_dim, )))
decoder.add(Activation('leaky_relu'))
decoder.add(BatchNormalization(momentum=0.8))
decoder.add(Dense(512))
decoder.add(Activation('leaky_relu'))
decoder.add(BatchNormalization(momentum=0.8))
decoder.add(Dense(self.img_dim))
decoder.add(Activation('tanh'))
return decoder
def train(self, n_epochs, batch_size=128, save_interval=50):
mnist = fetch_mldata('MNIST original')
X = mnist.data
y = mnist.target
# Rescale [-1, 1]
X = (X.astype(np.float32) - 127.5) / 127.5
for epoch in range(n_epochs):
# Select a random half batch of images
idx = np.random.randint(0, X.shape[0], batch_size)
imgs = X[idx]
# Train the Autoencoder
loss, _ = self.autoencoder.train_on_batch(imgs, imgs)
# Display the progress
print("%d [D loss: %f]" % (epoch, loss))
# If at save interval => save generated image samples
if epoch % save_interval == 0:
self.save_imgs(epoch, X)
def save_imgs(self, epoch, X):
r, c = 5, 5 # Grid size
# Select a random half batch of images
idx = np.random.randint(0, X.shape[0], r * c)
imgs = X[idx]
# Generate images and reshape to image shape
gen_imgs = self.autoencoder.predict(imgs).reshape(
(-1, self.img_rows, self.img_cols))
# Rescale images 0 - 1
gen_imgs = 0.5 * gen_imgs + 0.5
fig, axs = plt.subplots(r, c)
plt.suptitle("Autoencoder")
cnt = 0
for i in range(r):
for j in range(c):
axs[i, j].imshow(gen_imgs[cnt, :, :], cmap='gray')
axs[i, j].axis('off')
cnt += 1
fig.savefig("ae_%d.png" % epoch)
plt.close()
def main():
optimizer = Adadelta()
#-----
# MLP
#-----
data = datasets.load_digits()
X = data.data
y = data.target
# Convert to one-hot encoding
y = to_categorical(y.astype("int"))
n_samples = np.shape(X)
n_hidden = 512
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.4,
seed=1)
clf = NeuralNetwork(optimizer=optimizer,
loss=CrossEntropy,
validation_data=(X_test, y_test))
clf.add(Dense(n_hidden, input_shape=(8 * 8, )))
clf.add(Activation('leaky_relu'))
clf.add(Dense(n_hidden))
clf.add(Activation('leaky_relu'))
clf.add(Dropout(0.25))
clf.add(Dense(n_hidden))
clf.add(Activation('leaky_relu'))
clf.add(Dropout(0.25))
clf.add(Dense(n_hidden))
clf.add(Activation('leaky_relu'))
clf.add(Dropout(0.25))
clf.add(Dense(10))
clf.add(Activation('softmax'))
print()
clf.summary(name="MLP")
train_err, val_err = clf.fit(X_train, y_train, n_epochs=50, batch_size=256)
# Training and validation error plot
n = len(train_err)
training, = plt.plot(range(n), train_err, label="Training Error")
validation, = plt.plot(range(n), val_err, label="Validation Error")
plt.legend(handles=[training, validation])
plt.title("Error Plot")
plt.ylabel('Error')
plt.xlabel('Iterations')
plt.show()
# Predict labels of the test data
y_pred = np.argmax(clf.predict(X_test), axis=1)
y_test = np.argmax(y_test, axis=1)
accuracy = accuracy_score(y_test, y_pred)
print("Accuracy:", accuracy)
# Reduce dimension to 2D using PCA and plot the results
Plot().plot_in_2d(X_test,
y_pred,
title="Multilayer Perceptron",
accuracy=accuracy,
legend_labels=range(10))
def main():
#----------
# Conv Net
#----------
optimizer = Adam()
data = datasets.load_digits()
X = data.data
y = data.target
# Convert to one-hot encoding
y = to_categorical(y.astype("int"))
n_samples = np.shape(X)
n_hidden = 512
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.4, seed=1)
# Reshape X to (n_samples, channels, height, width)
X_train = X_train.reshape((-1,1,8,8))
X_test = X_test.reshape((-1,1,8,8))
clf = NeuralNetwork(optimizer=optimizer,
loss=CrossEntropy,
validation_data=(X_test, y_test))
clf.add(Conv2D(n_filters=16, filter_shape=(3,3), input_shape=(1,8,8), padding='same'))
clf.add(Activation('relu'))
clf.add(Dropout(0.25))
clf.add(BatchNormalization())
clf.add(Conv2D(n_filters=32, filter_shape=(3,3), padding='same'))
clf.add(Activation('relu'))
clf.add(Dropout(0.25))
clf.add(BatchNormalization())
clf.add(Flatten())
clf.add(Dense(256))
clf.add(Activation('relu'))
clf.add(Dropout(0.4))
clf.add(BatchNormalization())
clf.add(Dense(10))
clf.add(Activation('softmax'))
print ()
clf.summary(name="ConvNet")
train_err, val_err = clf.fit(X_train, y_train, n_epochs=50, batch_size=256)
# Training and validation error plot
n = len(train_err)
training, = plt.plot(range(n), train_err, label="Training Error")
validation, = plt.plot(range(n), val_err, label="Validation Error")
plt.legend(handles=[training, validation])
plt.title("Error Plot")
plt.ylabel('Error')
plt.xlabel('Iterations')
plt.show()
_, accuracy = clf.test_on_batch(X_test, y_test)
print ("Accuracy:", accuracy)
y_pred = np.argmax(clf.predict(X_test), axis=1)
X_test = X_test.reshape(-1, 8*8)
# Reduce dimension to 2D using PCA and plot the results
Plot().plot_in_2d(X_test, y_pred, title="Convolutional Neural Network", accuracy=accuracy, legend_labels=range(10))
def build_discriminator(self, optimizer, loss_function):
model = NeuralNetwork(optimizer=optimizer, loss=loss_function)
model.add(Dense(512, input_shape=(self.img_dim,)))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.5))
model.add(Dense(256))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.5))
model.add(Dense(2))
model.add(Activation('softmax'))
return model
def build_generator(self, optimizer, loss_function):
model = NeuralNetwork(optimizer=optimizer, loss=loss_function)
model.add(Dense(256, input_shape=(self.latent_dim,)))
model.add(Activation('leaky_relu'))
model.add(BatchNormalization(momentum=0.8))
model.add(Dense(512))
model.add(Activation('leaky_relu'))
model.add(BatchNormalization(momentum=0.8))
model.add(Dense(1024))
model.add(Activation('leaky_relu'))
model.add(BatchNormalization(momentum=0.8))
model.add(Dense(self.img_dim))
model.add(Activation('tanh'))
return model
class GAN():
"""A Generative Adversarial Network with deep fully-connected neural nets as
Generator and Discriminator.
Training Data: MNIST Handwritten Digits (28x28 images)
"""
def __init__(self):
self.img_rows = 28
self.img_cols = 28
self.img_dim = self.img_rows * self.img_cols
self.latent_dim = 100
optimizer = Adam(learning_rate=0.0002, b1=0.5)
loss_function = CrossEntropy
# Build the discriminator
self.discriminator = self.build_discriminator(optimizer, loss_function)
# Build the generator
self.generator = self.build_generator(optimizer, loss_function)
# Build the combined model
self.combined = NeuralNetwork(optimizer=optimizer, loss=loss_function)
self.combined.layers.extend(self.generator.layers)
self.combined.layers.extend(self.discriminator.layers)
print ()
self.generator.summary(name="Generator")
self.discriminator.summary(name="Discriminator")
def build_generator(self, optimizer, loss_function):
model = NeuralNetwork(optimizer=optimizer, loss=loss_function)
model.add(Dense(256, input_shape=(self.latent_dim,)))
model.add(Activation('leaky_relu'))
model.add(BatchNormalization(momentum=0.8))
model.add(Dense(512))
model.add(Activation('leaky_relu'))
model.add(BatchNormalization(momentum=0.8))
model.add(Dense(1024))
model.add(Activation('leaky_relu'))
model.add(BatchNormalization(momentum=0.8))
model.add(Dense(self.img_dim))
model.add(Activation('tanh'))
return model
def build_discriminator(self, optimizer, loss_function):
model = NeuralNetwork(optimizer=optimizer, loss=loss_function)
model.add(Dense(512, input_shape=(self.img_dim,)))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.5))
model.add(Dense(256))
model.add(Activation('leaky_relu'))
model.add(Dropout(0.5))
model.add(Dense(2))
model.add(Activation('softmax'))
return model
def train(self, n_epochs, batch_size=128, save_interval=50):
mnist = fetch_mldata('MNIST original')
X = mnist.data
y = mnist.target
# Rescale [-1, 1]
X = (X.astype(np.float32) - 127.5) / 127.5
half_batch = int(batch_size / 2)
for epoch in range(n_epochs):
# ---------------------
# Train Discriminator
# ---------------------
self.discriminator.set_trainable(True)
# Select a random half batch of images
idx = np.random.randint(0, X.shape[0], half_batch)
imgs = X[idx]
# Sample noise to use as generator input
noise = np.random.normal(0, 1, (half_batch, self.latent_dim))
# Generate a half batch of images
gen_imgs = self.generator.predict(noise)
# Valid = [1, 0], Fake = [0, 1]
valid = np.concatenate((np.ones((half_batch, 1)), np.zeros((half_batch, 1))), axis=1)
fake = np.concatenate((np.zeros((half_batch, 1)), np.ones((half_batch, 1))), axis=1)
# Train the discriminator
d_loss_real, d_acc_real = self.discriminator.train_on_batch(imgs, valid)
d_loss_fake, d_acc_fake = self.discriminator.train_on_batch(gen_imgs, fake)
d_loss = 0.5 * (d_loss_real + d_loss_fake)
d_acc = 0.5 * (d_acc_real + d_acc_fake)
# ---------------------
# Train Generator
# ---------------------
# We only want to train the generator for the combined model
self.discriminator.set_trainable(False)
# Sample noise and use as generator input
noise = np.random.normal(0, 1, (batch_size, self.latent_dim))
# The generator wants the discriminator to label the generated samples as valid
valid = np.concatenate((np.ones((batch_size, 1)), np.zeros((batch_size, 1))), axis=1)
# Train the generator
g_loss, g_acc = self.combined.train_on_batch(noise, valid)
# Display the progress
print ("%d [D loss: %f, acc: %.2f%%] [G loss: %f, acc: %.2f%%]" % (epoch, d_loss, 100*d_acc, g_loss, 100*g_acc))
# If at save interval => save generated image samples
if epoch % save_interval == 0:
self.save_imgs(epoch)
def save_imgs(self, epoch):
r, c = 5, 5 # Grid size
noise = np.random.normal(0, 1, (r * c, self.latent_dim))
# Generate images and reshape to image shape
gen_imgs = self.generator.predict(noise).reshape((-1, self.img_rows, self.img_cols))
# Rescale images 0 - 1
gen_imgs = 0.5 * gen_imgs + 0.5
fig, axs = plt.subplots(r, c)
plt.suptitle("Generative Adversarial Network")
cnt = 0
for i in range(r):
for j in range(c):
axs[i,j].imshow(gen_imgs[cnt,:,:], cmap='gray')
axs[i,j].axis('off')
cnt += 1
fig.savefig("mnist_%d.png" % epoch)
plt.close()
def main():
optimizer = Adam()
def gen_mult_ser(nums):
""" Method which generates multiplication series """
X = np.zeros([nums, 10, 61], dtype=float)
y = np.zeros([nums, 10, 61], dtype=float)
for i in range(nums):
start = np.random.randint(2, 7)
mult_ser = np.linspace(start, start * 10, num=10, dtype=int)
X[i] = to_categorical(mult_ser, n_col=61)
y[i] = np.roll(X[i], -1, axis=0)
y[:, -1, 1] = 1 # Mark endpoint as 1
return X, y
def gen_num_seq(nums):
""" Method which generates sequence of numbers """
X = np.zeros([nums, 10, 20], dtype=float)
y = np.zeros([nums, 10, 20], dtype=float)
for i in range(nums):
start = np.random.randint(0, 10)
num_seq = np.arange(start, start + 10)
X[i] = to_categorical(num_seq, n_col=20)
y[i] = np.roll(X[i], -1, axis=0)
y[:, -1, 1] = 1 # Mark endpoint as 1
return X, y
X, y = gen_mult_ser(3000)
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.4)
# Model definition
clf = NeuralNetwork(optimizer=optimizer,
loss=CrossEntropy)
clf.add(RNN(10, activation="tanh", bptt_trunc=5, input_shape=(10, 61)))
clf.add(Activation('softmax'))
clf.summary("RNN")
# Print a problem instance and the correct solution
tmp_X = np.argmax(X_train[0], axis=1)
tmp_y = np.argmax(y_train[0], axis=1)
print("Number Series Problem:")
print("X = [" + " ".join(tmp_X.astype("str")) + "]")
print("y = [" + " ".join(tmp_y.astype("str")) + "]")
print()
train_err, _ = clf.fit(X_train, y_train, n_epochs=500, batch_size=512)
# Predict labels of the test data
y_pred = np.argmax(clf.predict(X_test), axis=2)
y_test = np.argmax(y_test, axis=2)
print()
print("Results:")
for i in range(5):
# Print a problem instance and the correct solution
tmp_X = np.argmax(X_test[i], axis=1)
tmp_y1 = y_test[i]
tmp_y2 = y_pred[i]
print("X = [" + " ".join(tmp_X.astype("str")) + "]")
print("y_true = [" + " ".join(tmp_y1.astype("str")) + "]")
print("y_pred = [" + " ".join(tmp_y2.astype("str")) + "]")
print()
accuracy = np.mean(accuracy_score(y_test, y_pred))
print("Accuracy:", accuracy)
training = plt.plot(range(500), train_err, label="Training Error")
plt.title("Error Plot")
plt.ylabel('Training Error')
plt.xlabel('Iterations')
plt.show()
def main():
# define the model
components = 3
optimizer = Adam()
loss = MdnLoss(num_components=components, output_dim=1)
clf = NeuralNetwork(optimizer=optimizer, loss=loss)
clf.add(Dense(n_units=26, input_shape=(1, )))
clf.add(Activation('tanh'))
clf.add(
MDN(input_shape=(26, ), output_shape=(1, ), num_components=components))
clf.summary(name="MDN")
# generate 1D regression data (Bishop book, page 273).
# Note: P(y|x) is not a nice distribution.
# (e.g.) it has three modes for x ~= 0.5
N = 225
X = np.linspace(0, 1, N)
Y = X + 0.3 * np.sin(2 * 3.1415926 * X) + np.random.uniform(-0.1, 0.1, N)
X, Y = Y, X
nb = N # full_batch
xbatch = np.reshape(X[:nb], (nb, 1))
ybatch = np.reshape(Y[:nb], (nb, 1))
train_err, val_err = clf.fit(xbatch,
ybatch,
n_epochs=int(4e3),
batch_size=N)
plt.plot(train_err, label="Training Error")
plt.title("Error Plot")
plt.ylabel('Error')
plt.xlabel('Iterations')
plt.show()
# utility function for creating contour plot of the predictions
n = 15
xx = np.linspace(0, 1, n)
yy = np.linspace(0, 1, n)
xm, ym = np.meshgrid(xx, yy)
loss, acc = clf.test_on_batch(xm.reshape(xm.size, 1),
ym.reshape(ym.size, 1))
ypred = clf.loss_function.ypred
plt.figure(figsize=(10, 10))
plt.scatter(X, Y, color='g')
plt.contour(xm,
ym,
np.reshape(ypred, (n, n)),
levels=np.linspace(ypred.min(), ypred.max(), 20))
plt.xlabel('x')
plt.ylabel('y')
plt.title('{}-component Gaussian Mixture Model for '
'P(y|x)'.format(components))
plt.show()
def main():
#----------
# Conv Net
#----------
optimizer = Adam()
data = datasets.load_digits()
X = data.data
y = data.target
# Convert to one-hot encoding
y = to_categorical(y.astype("int"))
X_train, X_test, y_train, y_test = train_test_split(X,
y,
test_size=0.4,
seed=1)
# Reshape X to (n_samples, channels, height, width)
X_train = X_train.reshape((-1, 1, 8, 8))
X_test = X_test.reshape((-1, 1, 8, 8))
clf = NeuralNetwork(optimizer=optimizer,
loss=CrossEntropy,
validation_data=(X_test, y_test))
clf.add(
Conv2D(n_filters=16,
filter_shape=(3, 3),
stride=1,
input_shape=(1, 8, 8),
padding='same'))
clf.add(Activation('relu'))
clf.add(Dropout(0.25))
clf.add(BatchNormalization())
clf.add(Conv2D(n_filters=32, filter_shape=(3, 3), stride=1,
padding='same'))
clf.add(Activation('relu'))
clf.add(Dropout(0.25))
clf.add(BatchNormalization())
clf.add(Flatten())
clf.add(Dense(256))
clf.add(Activation('relu'))
clf.add(Dropout(0.4))
clf.add(BatchNormalization())
clf.add(Dense(10))
clf.add(Activation('softmax'))
print()
clf.summary(name="ConvNet")
train_err, val_err = clf.fit(X_train, y_train, n_epochs=50, batch_size=256)
# Training and validation error plot
n = len(train_err)
training, = plt.plot(range(n), train_err, label="Training Error")
validation, = plt.plot(range(n), val_err, label="Validation Error")
plt.legend(handles=[training, validation])
plt.title("Error Plot")
plt.ylabel('Error')
plt.xlabel('Iterations')
plt.show()
_, accuracy = clf.test_on_batch(X_test, y_test)
print("Accuracy:", accuracy)
y_pred = np.argmax(clf.predict(X_test), axis=1)
X_test = X_test.reshape(-1, 8 * 8)
# Reduce dimension to 2D using PCA and plot the results
Plot().plot_in_2d(X_test,
y_pred,
title="Convolutional Neural Network",
accuracy=accuracy,
legend_labels=range(10))
def main():
optimizer = Adam()
#-----
# MLP
#-----
data = datasets.load_digits()
X = data.data
y = data.target
# Convert to one-hot encoding
y = to_categorical(y.astype("int"))
n_samples, n_features = X.shape
n_hidden = 512
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.4, seed=1)
clf = NeuralNetwork(optimizer=optimizer,
loss=CrossEntropy,
validation_data=(X_test, y_test))
clf.add(Dense(n_hidden, input_shape=(n_features,)))
clf.add(Activation('leaky_relu'))
clf.add(Dense(n_hidden))
clf.add(Activation('leaky_relu'))
clf.add(Dropout(0.25))
clf.add(Dense(n_hidden))
clf.add(Activation('leaky_relu'))
clf.add(Dropout(0.25))
clf.add(Dense(n_hidden))
clf.add(Activation('leaky_relu'))
clf.add(Dropout(0.25))
clf.add(Dense(10))
clf.add(Activation('softmax'))
print ()
clf.summary(name="MLP")
train_err, val_err = clf.fit(X_train, y_train, n_epochs=50, batch_size=256)
# Training and validation error plot
n = len(train_err)
training, = plt.plot(range(n), train_err, label="Training Error")
validation, = plt.plot(range(n), val_err, label="Validation Error")
plt.legend(handles=[training, validation])
plt.title("Error Plot")
plt.ylabel('Error')
plt.xlabel('Iterations')
plt.show()
_, accuracy = clf.test_on_batch(X_test, y_test)
print ("Accuracy:", accuracy)
# Reduce dimension to 2D using PCA and plot the results
y_pred = np.argmax(clf.predict(X_test), axis=1)
Plot().plot_in_2d(X_test, y_pred, title="Multilayer Perceptron", accuracy=accuracy, legend_labels=range(10))
def model(n_inputs, n_outputs):
clf = NeuralNetwork(optimizer=Adam(), loss=SquareLoss)
clf.add(Dense(64, input_shape=(n_inputs, )))
clf.add(Activation('relu'))
clf.add(Dense(n_outputs))
return clf