def load_and_train_highway_network_initial(model_name, images_x, labels_y):
images_x = np.reshape(images_x, (-1, 192, 256, 1))
network = input_data(shape=[None, 192, 256, 1], name='input')
for i in range(5):
for j in [3, 2, 1]:
network = highway_conv_2d(network, 16, j, activation='elu')
network = max_pool_2d(network, 2)
network = batch_normalization(network)
network = fully_connected(network, 128, activation='elu')
network = fully_connected(network, 256, activation='elu')
network = fully_connected(network, 3, activation='softmax')
network = regression(network,
optimizer='adam',
learning_rate=0.01,
loss='categorical_crossentropy',
name='target')
model = tflearn.DNN(network, tensorboard_verbose=0)
model.fit(images_x,
labels_y,
n_epoch=40,
show_metric=True,
run_id='Q_values_highway')
model.save(model_name)
def ConvHighway1(network):
for i in range(3):
for j in [3, 2, 1]:
network = highway_conv_2d(network, 16, j, activation='elu')
network = max_pool_2d(network, 2)
network = batch_normalization(network)
network = fully_connected(network, 128, activation='elu')
network = fully_connected(network, 256, activation='elu')
network = fully_connected(network, output_dim, activation='softmax')
return network
def ConvHighway1(network, scale=False):
if scale is True:
network = scale(network)
for i in range(3):
for j in [3, 2, 1]:
network = highway_conv_2d(network, 16, j, activation='elu')
network = max_pool_2d(network, 2)
network = batch_normalization(network)
network = fully_connected(network, 128, activation='elu')
network = fully_connected(network, 256, activation='elu')
network = fully_connected(network, output_dim, activation='sigmoid')
return network
def network():
"""
Maciej A. Czyzewski @maciejczyzewski
PAMG - experimental neural network for regular structures
"""
# input
net = input_data(shape=[None, 21, 21, 1], name='input')
# H(2)
for i in range(2):
for j in [3, 2, 1]:
net = highway_conv_2d(net, 16, j, activation='elu')
net = max_pool_2d(net, 2)
net = batch_normalization(net)
# 2D(32)
net = conv_2d(net, 32, 3, activation='relu', regularizer="L2")
net = max_pool_2d(net, 2)
net = local_response_normalization(net)
# 2D(64)
net = conv_2d(net, 64, 3, activation='leaky_relu', regularizer="L2")
net = max_pool_2d(net, 3)
net = local_response_normalization(net)
# 2D(128)
net = conv_2d(net, 128, 3, activation='relu6', regularizer="L2")
net = max_pool_2d(net, 4)
net = local_response_normalization(net)
# F(128)
net = fully_connected(net, 128, activation='elu')
net = dropout(net, 0.5)
# F(256)
net = fully_connected(net, 256, activation='tanh')
net = dropout(net, 0.5)
# output
net = fully_connected(net, 2, activation='softmax')
return regression(net,
optimizer='adam',
learning_rate=0.003,
loss='categorical_crossentropy',
name='target')
def classifyHighway():
# ================================
# Building convolutional network
# ================================
network = input_data(shape=[None, buff_size, num_buffs, 2], name='input')
# highway convolutions with pooling and dropout
for i in range(2):
for j in [4, 6, 12]:
# network = dropout(network, 0.75)
# https://github.com/tflearn/tflearn/blob/2faad812dc35e08457dc6bd86b15392446cffd87/tflearn/layers/conv.py#L1346
network = highway_conv_2d(network, 4, j, activation='leaky_relu')
# https://github.com/tflearn/tflearn/blob/2faad812dc35e08457dc6bd86b15392446cffd87/tflearn/layers/conv.py#L266
network = max_pool_2d(network, 4)
# https://github.com/tflearn/tflearn/blob/2faad812dc35e08457dc6bd86b15392446cffd87/tflearn/layers/normalization.py#L20
network = batch_normalization(network)
# https://github.com/tflearn/tflearn/blob/51399601c1a4f305db894b871baf743baa15ea00/tflearn/layers/core.py#L96
network = fully_connected(network, 128, activation='prelu')
network = fully_connected(network, 32, activation='elu')
network = fully_connected(network, len(library), activation='softmax')
# https://github.com/tflearn/tflearn/blob/4ba8c8d78bf1bbdfc595bf547bad30580cb4c20b/tflearn/layers/estimator.py#L14
network = regression(network,
optimizer='adam',
learning_rate=0.005,
loss='categorical_crossentropy',
name='target')
print("Training")
# Training
# https://github.com/tflearn/tflearn/blob/66c0c9c67b0472cbdc85bae0beb7992fa008480e/tflearn/models/dnn.py#L10
model = tflearn.DNN(network, tensorboard_verbose=3)
# https://github.com/tflearn/tflearn/blob/66c0c9c67b0472cbdc85bae0beb7992fa008480e/tflearn/models/dnn.py#L89
model.fit(X,
Y,
n_epoch=15,
validation_set=(testX, testY),
show_metric=True,
run_id='convnet_highway_dsp')
# Validation
pred = model.predict(valX)
val_acc = compare(pred, valY)
print("Validation accuracy : ", val_acc)
def __init__(self, learning_rate):
# Real-time data preprocessing
img_prep = ImagePreprocessing()
img_prep.add_featurewise_zero_center()
img_prep.add_featurewise_stdnorm()
# Convolutional network building
network = input_data(shape=[None, 32, 32, 3],data_preprocessing=img_prep)
#highway convolutions with pooling and dropout
for i in range(3):
for j in [3, 2, 1]:
network = highway_conv_2d(network, 16, j, activation='elu')
network = max_pool_2d(network, 2)
network = batch_normalization(network)
network = fully_connected(network, 128, activation='elu')
network = dropout(network, 0.5)
network = fully_connected(network, 256, activation='elu')
network = fully_connected(network, 10, activation='softmax')
network = regression(network, optimizer='adam', learning_rate= learning_rate,
loss='categorical_crossentropy')
self.network = network
from tflearn.layers.conv import highway_conv_2d, max_pool_2d
from tflearn.layers.normalization import local_response_normalization, batch_normalization
from tflearn.layers.estimator import regression
# Data loading and preprocessing
import tflearn.datasets.mnist as mnist
X, Y, testX, testY = mnist.load_data(one_hot=True)
X = X.reshape([-1, 28, 28, 1])
testX = testX.reshape([-1, 28, 28, 1])
# Building convolutional network
network = input_data(shape=[None, 28, 28, 1], name='input')
#highway convolutions with pooling and dropout
for i in range(3):
for j in [3, 2, 1]:
network = highway_conv_2d(network, 16, j, activation='elu')
network = max_pool_2d(network, 2)
network = batch_normalization(network)
network = fully_connected(network, 128, activation='elu')
network = fully_connected(network, 256, activation='elu')
network = fully_connected(network, 10, activation='softmax')
network = regression(network,
optimizer='adam',
learning_rate=0.01,
loss='categorical_crossentropy',
name='target')
# Training
model = tflearn.DNN(network, tensorboard_verbose=0)
model.fit(X,
from tflearn.layers.core import input_data, dropout, fully_connected
from tflearn.layers.conv import highway_conv_2d, max_pool_2d
from tflearn.layers.normalization import local_response_normalization, batch_normalization
from tflearn.layers.estimator import regression
from deepdsp.conf import conf
network = input_data(shape=[None, conf["buff_size"], conf["num_buffs"], 4], name='input')
# highway convolutions with pooling and dropout
for i in range(2):
for j in [4, 6, 12]:
network = dropout(network, 0.95)
# https://github.com/tflearn/tflearn/blob/2faad812dc35e08457dc6bd86b15392446cffd87/tflearn/layers/conv.py#L1346
network = highway_conv_2d(network, 4, j, activation='leaky_relu')
# https://github.com/tflearn/tflearn/blob/2faad812dc35e08457dc6bd86b15392446cffd87/tflearn/layers/conv.py#L266
# network = max_pool_2d(network, 4)
network = max_pool_2d(network, int(8/(i+1)))
# https://github.com/tflearn/tflearn/blob/2faad812dc35e08457dc6bd86b15392446cffd87/tflearn/layers/normalization.py#L20
network = batch_normalization(network)
# https://github.com/tflearn/tflearn/blob/51399601c1a4f305db894b871baf743baa15ea00/tflearn/layers/core.py#L96
network = fully_connected(network, 150, activation='prelu')
network = fully_connected(network, 32, activation='elu')
network = fully_connected(network, len(conf["classes"]), activation='softmax')
# https://github.com/tflearn/tflearn/blob/4ba8c8d78bf1bbdfc595bf547bad30580cb4c20b/tflearn/layers/estimator.py#L14
network = regression(network, optimizer='adam', learning_rate=0.01,
loss='categorical_crossentropy', name='target')
from tflearn.layers.core import input_data, dropout, fully_connected
from tflearn.layers.conv import highway_conv_2d, max_pool_2d
from tflearn.layers.normalization import local_response_normalization, batch_normalization
from tflearn.layers.estimator import regression
# Data loading and preprocessing
import tflearn.datasets.mnist as mnist
X, Y, testX, testY = mnist.load_data(one_hot=True)
X = X.reshape([-1, 28, 28, 1])
testX = testX.reshape([-1, 28, 28, 1])
# Building convolutional network
network = input_data(shape=[None, 28, 28, 1], name='input')
#highway convolutions with pooling and dropout
for i in range(3):
for j in [3, 2, 1]:
network = highway_conv_2d(network, 16, j, activation='elu')
network = max_pool_2d(network, 2)
network = batch_normalization(network)
network = fully_connected(network, 128, activation='elu')
network = fully_connected(network, 256, activation='elu')
network = fully_connected(network, 10, activation='softmax')
network = regression(network, optimizer='adam', learning_rate=0.01,
loss='categorical_crossentropy', name='target')
# Training
model = tflearn.DNN(network, tensorboard_verbose=0)
model.fit(X, Y, n_epoch=20, validation_set=(testX, testY),
show_metric=True, run_id='convnet_highway_mnist')
