def __init__(self, nclass, h, w, c):
layers = []
identityblk = IdentityBlock(input_channels=c,
input_shape=[h, w],
nlayers=10)
layers.append(identityblk)
layers.append(
Conv2D(input_channels=c,
num_filters=16,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(1, 1), kernel_size=(3, 3))
layers.append(BatchNormalization(input_shape=[h, w, 16]))
denseblk = DenseBlock(input_channels=16,
input_shape=[h, w],
growth_rate=4,
nlayers=4)
layers.append(denseblk)
layers.append(
Conv2D(input_channels=denseblk.output_channels,
num_filters=32,
kernel_size=(3, 3),
stride=(2, 2),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(2, 2), kernel_size=(3, 3))
layers.append(Dropout(0.5))
layers.append(
Conv2D(input_channels=32,
num_filters=nclass,
kernel_size=(1, 1),
stride=(1, 1),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(1, 1), kernel_size=(1, 1))
layers.append(BatchNormalization(input_shape=[h, w, nclass]))
layers.append(
AvgPooling(poolsize=(h, w), stride=(1, 1), padding='VALID'))
layers.append(Flatten())
layers.append(Softmax())
self.startnode = tg.StartNode(input_vars=[None])
model_hn = tg.HiddenNode(prev=[self.startnode], layers=layers)
self.endnode = tg.EndNode(prev=[model_hn])
def model(nclass, h, w, c):
with tf.name_scope('Cifar10AllCNN'):
seq = tg.Sequential()
seq.add(Conv2D(input_channels=c, num_filters=96, kernel_size=(3, 3), stride=(1, 1), padding='SAME'))
seq.add(RELU())
seq.add(TFBatchNormalization(name='b1'))
h, w = same(in_height=h, in_width=w, stride=(1,1), kernel_size=(3,3))
seq.add(Conv2D(input_channels=96, num_filters=96, kernel_size=(3, 3), stride=(1, 1), padding='SAME'))
seq.add(RELU())
h, w = same(in_height=h, in_width=w, stride=(1,1), kernel_size=(3,3))
seq.add(Dropout(0.5))
seq.add(Conv2D(input_channels=96, num_filters=96, kernel_size=(3, 3), stride=(2, 2), padding='SAME'))
seq.add(RELU())
seq.add(TFBatchNormalization(name='b3'))
h, w = same(in_height=h, in_width=w, stride=(2,2), kernel_size=(3,3))
seq.add(Conv2D(input_channels=96, num_filters=192, kernel_size=(3, 3), stride=(1, 1), padding='SAME'))
seq.add(RELU())
h, w = same(in_height=h, in_width=w, stride=(1,1), kernel_size=(3,3))
seq.add(Dropout(0.5))
seq.add(Conv2D(input_channels=192, num_filters=192, kernel_size=(3, 3), stride=(1, 1), padding='SAME'))
seq.add(RELU())
seq.add(TFBatchNormalization(name='b5'))
h, w = same(in_height=h, in_width=w, stride=(1,1), kernel_size=(3,3))
seq.add(Conv2D(input_channels=192, num_filters=192, kernel_size=(3, 3), stride=(2, 2), padding='SAME'))
seq.add(RELU())
h, w = same(in_height=h, in_width=w, stride=(2,2), kernel_size=(3,3))
seq.add(Dropout(0.5))
seq.add(Conv2D(input_channels=192, num_filters=192, kernel_size=(3, 3), stride=(1, 1), padding='SAME'))
seq.add(RELU())
seq.add(TFBatchNormalization(name='b7'))
h, w = same(in_height=h, in_width=w, stride=(1,1), kernel_size=(3,3))
seq.add(Conv2D(input_channels=192, num_filters=192, kernel_size=(1, 1), stride=(1, 1), padding='SAME'))
seq.add(RELU())
h, w = same(in_height=h, in_width=w, stride=(1,1), kernel_size=(1,1))
seq.add(Dropout(0.5))
seq.add(Conv2D(input_channels=192, num_filters=nclass, kernel_size=(1, 1), stride=(1, 1), padding='SAME'))
seq.add(RELU())
seq.add(TFBatchNormalization(name='b9'))
h, w = same(in_height=h, in_width=w, stride=(1,1), kernel_size=(1,1))
seq.add(AvgPooling(poolsize=(h, w), stride=(1,1), padding='VALID'))
seq.add(Flatten())
seq.add(Softmax())
return seq
def model():
with tf.name_scope('MnistCNN'):
seq = tg.Sequential()
seq.add(
Conv2D(input_channels=1,
num_filters=32,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'))
h, w = same(in_height=28,
in_width=28,
stride=(1, 1),
kernel_size=(3, 3))
seq.add(BatchNormalization(input_shape=[h, w, 32]))
seq.add(RELU())
seq.add(MaxPooling(poolsize=(2, 2), stride=(2, 2), padding='SAME'))
h, w = same(in_height=h, in_width=w, stride=(2, 2), kernel_size=(2, 2))
seq.add(LRN())
seq.add(
Conv2D(input_channels=32,
num_filters=64,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'))
h, w = same(in_height=h, in_width=w, stride=(1, 1), kernel_size=(3, 3))
seq.add(BatchNormalization(input_shape=[h, w, 64]))
seq.add(RELU())
seq.add(MaxPooling(poolsize=(2, 2), stride=(2, 2), padding='SAME'))
h, w = same(in_height=h, in_width=w, stride=(2, 2), kernel_size=(2, 2))
seq.add(LRN())
seq.add(Flatten())
seq.add(Linear(int(h * w * 64), 128))
seq.add(BatchNormalization(input_shape=[128]))
seq.add(Tanh())
seq.add(Dropout(0.8))
seq.add(Linear(128, 256))
seq.add(BatchNormalization(input_shape=[256]))
seq.add(Tanh())
seq.add(Dropout(0.8))
seq.add(Linear(256, 10))
seq.add(Softmax())
return seq
def __init__(self, nclass, h, w, c):
layers = []
layers.append(
Conv2D(input_channels=c,
num_filters=96,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(1, 1), kernel_size=(3, 3))
layers.append(BatchNormalization(input_shape=[h, w, 96]))
layers.append(
Conv2D(input_channels=96,
num_filters=96,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(1, 1), kernel_size=(3, 3))
layers.append(Dropout(0.5))
layers.append(
Conv2D(input_channels=96,
num_filters=96,
kernel_size=(3, 3),
stride=(2, 2),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(2, 2), kernel_size=(3, 3))
layers.append(BatchNormalization(input_shape=[h, w, 96]))
layers.append(
Conv2D(input_channels=96,
num_filters=192,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(1, 1), kernel_size=(3, 3))
layers.append(Dropout(0.5))
layers.append(
Conv2D(input_channels=192,
num_filters=192,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(1, 1), kernel_size=(3, 3))
layers.append(BatchNormalization(input_shape=[h, w, 192]))
layers.append(
Conv2D(input_channels=192,
num_filters=192,
kernel_size=(3, 3),
stride=(2, 2),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(2, 2), kernel_size=(3, 3))
layers.append(Dropout(0.5))
layers.append(
Conv2D(input_channels=192,
num_filters=192,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(1, 1), kernel_size=(3, 3))
layers.append(BatchNormalization(input_shape=[h, w, 192]))
layers.append(
Conv2D(input_channels=192,
num_filters=192,
kernel_size=(1, 1),
stride=(1, 1),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(1, 1), kernel_size=(1, 1))
layers.append(Dropout(0.5))
layers.append(
Conv2D(input_channels=192,
num_filters=nclass,
kernel_size=(1, 1),
stride=(1, 1),
padding='SAME'))
layers.append(RELU())
h, w = same(in_height=h, in_width=w, stride=(1, 1), kernel_size=(1, 1))
layers.append(BatchNormalization(input_shape=[h, w, nclass]))
layers.append(
AvgPooling(poolsize=(h, w), stride=(1, 1), padding='VALID'))
layers.append(Flatten())
layers.append(Softmax())
self.startnode = tg.StartNode(input_vars=[None])
model_hn = tg.HiddenNode(prev=[self.startnode], layers=layers)
self.endnode = tg.EndNode(prev=[model_hn])
def classifier(X_ph, X_gen_ph, h, w):
with tf.variable_scope('Classifier'):
X_sn = tg.StartNode(input_vars=[X_ph])
X_gen_sn = tg.StartNode(input_vars=[X_gen_ph])
h1, w1 = same(in_height=h,
in_width=w,
stride=(1, 1),
kernel_size=(3, 3))
h2, w2 = same(in_height=h1,
in_width=w1,
stride=(2, 2),
kernel_size=(2, 2))
h3, w3 = same(in_height=h2,
in_width=w2,
stride=(1, 1),
kernel_size=(3, 3))
h4, w4 = same(in_height=h3,
in_width=w3,
stride=(2, 2),
kernel_size=(2, 2))
print('---', h, w)
X_hn = tg.HiddenNode(prev=[X_sn],
layers=[
Conv2D(input_channels=1,
num_filters=32,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'),
BatchNormalization(input_shape=[h1, w1, 32]),
RELU(),
MaxPooling(poolsize=(2, 2),
stride=(2, 2),
padding='SAME'),
LRN(),
Conv2D(input_channels=32,
num_filters=64,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'),
BatchNormalization(input_shape=[h3, w3, 64]),
RELU(),
MaxPooling(poolsize=(2, 2),
stride=(2, 2),
padding='SAME'),
Flatten(),
])
X_gen_hn = tg.HiddenNode(
prev=[X_gen_sn],
layers=[
Conv2D(input_channels=1,
num_filters=32,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'),
BatchNormalization(input_shape=[h1, w1, 32]),
RELU(),
MaxPooling(poolsize=(2, 2), stride=(2, 2), padding='SAME'),
LRN(),
Conv2D(input_channels=32,
num_filters=64,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'),
BatchNormalization(input_shape=[h3, w3, 64]),
RELU(),
MaxPooling(poolsize=(2, 2), stride=(2, 2), padding='SAME'),
Flatten(),
])
print('===', h4 * w4 * 64 * 2)
merge_hn = tg.HiddenNode(prev=[X_hn, X_gen_hn],
input_merge_mode=Concat(),
layers=[
Linear(h4 * w4 * 64 * 2, 100),
RELU(),
BatchNormalization(input_shape=[100]),
Linear(100, 1),
Sigmoid()
])
en = tg.EndNode(prev=[merge_hn])
graph = tg.Graph(start=[X_sn, X_gen_sn], end=[en])
y_train, = graph.train_fprop()
y_test, = graph.test_fprop()
return y_train, y_test
def discriminator_allconv(self):
if not self.generator_called:
raise Exception(
'self.generator() has to be called first before self.discriminator()'
)
scope = 'Discriminator'
with self.tf_graph.as_default():
with tf.name_scope(scope):
# h1, w1 = valid(self.h, self.w, kernel_size=(5,5), stride=(1,1))
# h2, w2 = valid(h1, w1, kernel_size=(5,5), stride=(2,2))
# h3, w3 = valid(h2, w2, kernel_size=(5,5), stride=(2,2))
# flat_dim = int(h3*w3*32)
dis_real_sn = tg.StartNode(input_vars=[self.real_ph])
# fake_ph = tf.placeholder('float32', [None, self.h, self.w, 1], name='fake')
# fake_sn = tg.StartNode(input_vars=[fake_ph])
h, w = same(in_height=self.h,
in_width=self.w,
stride=(1, 1),
kernel_size=(3, 3))
h, w = same(in_height=h,
in_width=w,
stride=(1, 1),
kernel_size=(3, 3))
h, w = same(in_height=h,
in_width=w,
stride=(2, 2),
kernel_size=(3, 3))
h, w = same(in_height=h,
in_width=w,
stride=(1, 1),
kernel_size=(3, 3))
h, w = same(in_height=h,
in_width=w,
stride=(1, 1),
kernel_size=(3, 3))
h, w = same(in_height=h,
in_width=w,
stride=(2, 2),
kernel_size=(3, 3))
h, w = same(in_height=h,
in_width=w,
stride=(1, 1),
kernel_size=(1, 1))
h, w = same(in_height=h,
in_width=w,
stride=(1, 1),
kernel_size=(3, 3))
h, w = same(in_height=h,
in_width=w,
stride=(1, 1),
kernel_size=(1, 1))
print('h, w', h, w)
print('===============')
# h, w = valid(in_height=h, in_width=w, stride=(1,1), kernel_size=(h,w))
disc_hn = tg.HiddenNode(
prev=[dis_real_sn, self.gen_hn],
layers=[
Dropout(0.2),
# TFBatchNormalization(name='b0'),
Conv2D(input_channels=self.c,
num_filters=96,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'),
LeakyRELU(),
TFBatchNormalization(name='b1'),
# Dropout(0.5),
Conv2D(input_channels=96,
num_filters=96,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'),
LeakyRELU(),
# TFBatchNormalization(name='b2'),
Dropout(0.5),
Conv2D(input_channels=96,
num_filters=96,
kernel_size=(3, 3),
stride=(2, 2),
padding='SAME'),
LeakyRELU(),
TFBatchNormalization(name='b3'),
# Dropout(0.5),
Conv2D(input_channels=96,
num_filters=192,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'),
LeakyRELU(),
# TFBatchNormalization(name='b4'),
Dropout(0.5),
Conv2D(input_channels=192,
num_filters=192,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'),
LeakyRELU(),
TFBatchNormalization(name='b5'),
# Dropout(0.5),
Conv2D(input_channels=192,
num_filters=192,
kernel_size=(3, 3),
stride=(2, 2),
padding='SAME'),
LeakyRELU(),
# TFBatchNormalization(name='b6'),
Dropout(0.5),
Conv2D(input_channels=192,
num_filters=192,
kernel_size=(3, 3),
stride=(1, 1),
padding='SAME'),
LeakyRELU(),
TFBatchNormalization(name='b7'),
# Dropout(0.5),
Conv2D(input_channels=192,
num_filters=192,
kernel_size=(1, 1),
stride=(1, 1),
padding='SAME'),
LeakyRELU(),
# TFBatchNormalization(name='b8'),
Dropout(0.5),
Conv2D(input_channels=192,
num_filters=self.nclass,
kernel_size=(1, 1),
stride=(1, 1),
padding='SAME'),
LeakyRELU(),
TFBatchNormalization(name='b9'),
# Dropout(0.5),
AvgPooling(poolsize=(h, w),
stride=(1, 1),
padding='VALID'),
Flatten(),
])
print('h,w', h, w)
print('==============')
class_hn = tg.HiddenNode(
prev=[disc_hn],
layers=[
Linear(self.nclass, self.nclass),
# Softmax()
])
judge_hn = tg.HiddenNode(
prev=[disc_hn],
layers=[
Linear(self.nclass, 1),
# Sigmoid()
])
real_class_en = tg.EndNode(prev=[class_hn])
real_judge_en = tg.EndNode(prev=[judge_hn])
fake_class_en = tg.EndNode(prev=[class_hn])
fake_judge_en = tg.EndNode(prev=[judge_hn])
graph = tg.Graph(start=[dis_real_sn],
end=[real_class_en, real_judge_en])
real_train = graph.train_fprop()
real_valid = graph.test_fprop()
graph = tg.Graph(start=[self.noise_sn, self.gen_real_sn],
end=[fake_class_en, fake_judge_en])
fake_train = graph.train_fprop()
fake_valid = graph.test_fprop()
dis_var_list = tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES,
scope=scope)
return self.real_ph, real_train, real_valid, fake_train, fake_valid, dis_var_list
def train():
batchsize = 64
learning_rate = 0.001
max_epoch = 10
X_train = np.random.rand(1000, 32, 32, 3)
M_train = np.random.rand(1000, 32, 32, 1)
X_valid = np.random.rand(1000, 32, 32, 3)
M_valid = np.random.rand(1000, 32, 32, 1)
X_ph = tf.placeholder('float32', [None, 32, 32, 3])
M_ph = tf.placeholder('float32', [None, 32, 32, 1])
h, w = 32, 32
model = tg.Sequential()
model.add(
Conv2D(input_channels=3,
num_filters=8,
kernel_size=(5, 5),
stride=(2, 2),
padding='SAME'))
h1, w1 = same(h, w, kernel_size=(5, 5), stride=(2, 2))
model.add(RELU())
model.add(
Conv2D(input_channels=8,
num_filters=16,
kernel_size=(5, 5),
stride=(2, 2),
padding='SAME'))
h2, w2 = same(h1, w1, kernel_size=(5, 5), stride=(2, 2))
model.add(RELU())
model.add(
Conv2D_Transpose(input_channels=16,
num_filters=8,
output_shape=(h1, w1),
kernel_size=(5, 5),
stride=(2, 2),
padding='SAME'))
model.add(RELU())
model.add(
Conv2D_Transpose(input_channels=8,
num_filters=1,
output_shape=(h, w),
kernel_size=(5, 5),
stride=(2, 2),
padding='SAME'))
model.add(RELU())
iter_model = tg.Sequential()
iter_model.add(
Conv2D(input_channels=1,
num_filters=8,
kernel_size=(5, 5),
stride=(2, 2),
padding='SAME'))
iter_model.add(RELU())
iter_model.add(
Conv2D_Transpose(input_channels=8,
num_filters=1,
output_shape=(h, w),
kernel_size=(5, 5),
stride=(2, 2),
padding='SAME'))
model.add(Iterative(sequential=iter_model, num_iter=10))
M_train_s = model.train_fprop(X_ph)
M_valid_s = model.test_fprop(X_ph)
train_mse = tf.reduce_mean((M_ph - M_train_s)**2)
valid_mse = tf.reduce_mean((M_ph - M_valid_s)**2)
data_train = tg.SequentialIterator(X_train, M_train, batchsize=batchsize)
data_valid = tg.SequentialIterator(X_valid, M_valid, batchsize=batchsize)
optimizer = tf.train.AdamOptimizer(learning_rate).minimize(train_mse)
with tf.Session() as sess:
init = tf.global_variables_initializer()
sess.run(init)
for epoch in range(max_epoch):
print('epoch:', epoch)
print('..training')
for X_batch, M_batch in data_train:
sess.run(optimizer, feed_dict={X_ph: X_batch, M_ph: M_batch})
print('..validating')
valid_mse_score = sess.run(valid_mse,
feed_dict={
X_ph: X_valid,
M_ph: M_valid
})
print('valid mse score:', valid_mse_score)
