def inference(x, num_output, wd, dropout_rate, is_training, transfer_mode= False):
with tf.variable_scope('conv1'):
network = common.spatialConvolution(x, 11, 4, 64, wd= wd)
network = common.batchNormalization(network, is_training= is_training)
network = tf.nn.relu (network)
#common.activation_summary(network)
network = common.maxPool(network, 3, 2)
with tf.variable_scope('conv2'):
network = common.spatialConvolution(network, 5, 1, 192, wd= wd)
network = common.batchNormalization(network, is_training= is_training)
network = tf.nn.relu(network)
#common.activation_summary(network)
network = common.maxPool(network, 3, 2)
with tf.variable_scope('conv3'):
network = common.spatialConvolution(network, 3, 1, 384, wd= wd)
network = common.batchNormalization(network, is_training= is_training)
network = tf.nn.relu(network)
#common.activation_summary(network)
with tf.variable_scope('conv4'):
network = common.spatialConvolution(network, 3, 1, 256, wd= wd)
network = common.batchNormalization(network, is_training= is_training)
network = tf.nn.relu(network)
with tf.variable_scope('conv5'):
network = common.spatialConvolution(network, 3, 1, 256, wd= wd)
network = common.batchNormalization(network, is_training= is_training)
network = tf.nn.relu(network)
network = common.maxPool(network, 3, 2)
network = common.flatten(network)
with tf.variable_scope('fc1'):
network = tf.nn.dropout(network, dropout_rate)
network = common.fullyConnected(network, 4096, wd= wd)
network = common.batchNormalization(network, is_training= is_training)
network = tf.nn.relu(network)
with tf.variable_scope('fc2'):
network = tf.nn.dropout(network, dropout_rate)
network = common.fullyConnected(network, 4096, wd= wd)
network = common.batchNormalization(network, is_training= is_training)
network = tf.nn.relu(network)
if not transfer_mode:
with tf.variable_scope('output'):
network = common.fullyConnected(network, num_output, wd= wd)
else:
with tf.variable_scope('transfer_output'):
network = common.fullyConnected(network, num_output, wd= wd)
return network
def getModel(x, num_chars, wd, is_training):
fc_weight_initializer = tf.truncated_normal_initializer(stddev=0.01)
batch_size = x.shape[0]
print('x in getModel', x) # Tensor("batch:0", shape=(32, ?, 26, 1), dtype=float32, device=/device:CPU:0)
with tf.variable_scope('dense1'):
# x = tf.squeeze(x)
x = tf.reshape(x, [-1, 26])
x = common.fullyConnected(x, 512, weight_initializer=fc_weight_initializer,
bias_initializer=tf.zeros_initializer, wd=wd)
x = tf.minimum(tf.nn.relu(x), 20)
print('x dense1', x) # (N*T, 512)
with tf.variable_scope('dense2'):
x = common.fullyConnected(x, 512, weight_initializer=fc_weight_initializer, bias_initializer=tf.zeros_initializer, wd=wd)
x = tf.minimum(tf.nn.relu(x), 20)
print('x dense2', x) # (N*T, 512)
with tf.variable_scope('dense3'):
x = common.fullyConnected(x, 512, weight_initializer=fc_weight_initializer, bias_initializer=tf.zeros_initializer, wd=wd)
x = tf.reshape(x, [-1, batch_size, 512])
x = tf.minimum(tf.nn.relu(x), 20)
print('!!! xin dense3: {}'.format(x)) # (?, 32, 4333)
with tf.variable_scope('birnn'):
x = common.biRNN(x=x, n_cell_dim=1024)
# Time major = True
print('!!!!!!x in birnn: {}'.format(x)) # (?, 512)
with tf.variable_scope('dense4'):
print('x start dense4', x)
x = common.fullyConnected(x, num_chars, weight_initializer=fc_weight_initializer, bias_initializer=tf.zeros_initializer, wd=wd)
x = tf.reshape(x, [-1, batch_size, num_chars])
print('!!! xin dense3: {}'.format(x)) # (?, 32, 4333)
# if not transfer_mode:
# with tf.variable_scope('output'):
# x = common.fullyConnected(x, num_chars, wd=wd)
# else:
# with tf.variable_scope('transfer_output'):
# x = common.fullyConnected(x, num_chars, wd=wd)
logging.info(x)
print('return x', x)
return x
def inference(x, num_output, wd, dropout_rate, is_training, transfer_mode= False, model_type= 'A'):
# Create tables describing VGG configurations A, B, D, E
if model_type == 'A':
config = [64, 'M', 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M']
elif model_type == 'B':
config = [64, 64, 'M', 128, 128, 'M', 256, 256, 'M', 512, 512, 'M', 512, 512, 'M']
elif model_type == 'D':
config = [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 'M', 512, 512, 512, 'M', 512, 512, 512, 'M']
elif model_type == 'E':
config = [64, 64, 'M', 128, 128, 'M', 256, 256, 256, 256, 'M', 512, 512, 512, 512, 'M', 512, 512, 512, 512, 'M']
else:
print('Unknown model type: ' + model_type + ' | Please specify a modelType A or B or D or E')
network= x
for k,v in enumerate(config):
if v == 'M':
network= common.maxPool(network, 2, 2)
else:
with tf.variable_scope('conv'+str(k)):
network = common.spatialConvolution(network, 3, 1, v, wd= wd)
network = tf.nn.relu(network)
network= common.flatten(network)
with tf.variable_scope('fc1'):
network = common.fullyConnected(network, 4096, wd= wd)
network = tf.nn.relu(network)
network = common.batchNormalization(network, is_training= is_training)
network = tf.nn.dropout(network, dropout_rate)
with tf.variable_scope('fc2'):
network = common.fullyConnected(network, 4096, wd= wd)
network = tf.nn.relu(network)
network = common.batchNormalization(network, is_training= is_training)
network = tf.nn.dropout(network, dropout_rate)
if not transfer_mode:
with tf.variable_scope('output'):
network = common.fullyConnected(network, num_output, wd= wd)
else:
with tf.variable_scope('transfer_output'):
network = common.fullyConnected(network, num_output, wd= wd)
return network
def getModel(x, num_output, K, stages, wd, is_training, transfer_mode=False):
with tf.variable_scope('conv1'):
x = common.spatialConvolution(x, 7, 2, 2 * K, wd=wd)
x = common.batchNormalization(x, is_training=is_training)
x = tf.nn.relu(x)
x = common.maxPool(x, 3, 2)
with tf.variable_scope('block1'):
x = block(x, stages[0], K, is_training=is_training, wd=wd)
with tf.variable_scope('trans1'):
x = transition(x, K, wd=wd, is_training=is_training)
with tf.variable_scope('block2'):
x = block(x, stages[1], K, is_training=is_training, wd=wd)
with tf.variable_scope('trans2'):
x = transition(x, K, wd=wd, is_training=is_training)
with tf.variable_scope('block3'):
x = block(x, stages[2], K, is_training=is_training, wd=wd)
with tf.variable_scope('trans3'):
x = transition(x, K, wd=wd, is_training=is_training)
with tf.variable_scope('block4'):
x = block(x, stages[3], K, is_training=is_training, wd=wd)
x = common.avgPool(x, 7, 1, padding='VALID')
x = common.flatten(x)
if not transfer_mode:
with tf.variable_scope('output'):
x = common.fullyConnected(x, num_output, wd=wd)
else:
with tf.variable_scope('transfer_output'):
x = common.fullyConnected(x, num_output, wd=wd)
return x
def inference(x, num_output, wd, is_training, transfer_mode= False):
with tf.variable_scope('block1'):
network = block(x, [11, 4, 96], wd, is_training)
network = common.maxPool(network, 3, 2)
with tf.variable_scope('block2'):
network = block(network, [5, 1, 256], wd, is_training)
network = common.maxPool(network, 3, 2)
with tf.variable_scope('block3'):
network = block(network, [3, 1, 384], wd, is_training)
network = common.maxPool(network, 3, 2)
with tf.variable_scope('block4'):
network = block(network, [3, 1, 1024], wd, is_training)
network = common.avgPool(network, 7, 1)
network = common.flatten(network)
if not transfer_mode:
with tf.variable_scope('output'):
network = common.fullyConnected(network, num_output, wd= wd)
else:
with tf.variable_scope('transfer_output'):
network = common.fullyConnected(network, num_output, wd= wd)
return network
def getModel(
x,
num_output,
wd,
is_training,
num_blocks=[3, 4, 6, 3], # defaults to 50-layer network
bottleneck=True):
conv_weight_initializer = tf.truncated_normal_initializer(stddev=0.1)
fc_weight_initializer = tf.truncated_normal_initializer(stddev=0.01)
with tf.variable_scope('scale1'):
x = common.spatialConvolution(
x, 7, 2, 64, weight_initializer=conv_weight_initializer, wd=wd)
x = common.batchNormalization(x, is_training=is_training)
x = tf.nn.relu(x)
with tf.variable_scope('scale2'):
x = common.maxPool(x, 3, 2)
x = common.resnetStack(x,
num_blocks[0],
1,
64,
bottleneck,
wd=wd,
is_training=is_training)
with tf.variable_scope('scale3'):
x = common.resnetStack(x,
num_blocks[1],
2,
128,
bottleneck,
wd=wd,
is_training=is_training)
with tf.variable_scope('scale4'):
x = common.resnetStack(x,
num_blocks[2],
2,
256,
bottleneck,
wd=wd,
is_training=is_training)
with tf.variable_scope('scale5'):
x = common.resnetStack(x,
num_blocks[3],
2,
512,
bottleneck,
wd=wd,
is_training=is_training)
# post-net
x = tf.reduce_mean(x, reduction_indices=[1, 2], name="avg_pool")
with tf.variable_scope('output'):
x = common.fullyConnected(x,
num_output,
weight_initializer=fc_weight_initializer,
bias_initializer=tf.zeros_initializer,
wd=wd)
return x
def inference(x,
num_output,
wd,
dropout_rate,
is_training,
transfer_mode=False):
with tf.variable_scope('features'):
with tf.variable_scope('conv1'):
network = common.spatialConvolution(x, 7, 2, 64, wd=wd)
network = common.batchNormalization(network,
is_training=is_training)
network = tf.nn.relu(network)
network = common.maxPool(network, 3, 2)
with tf.variable_scope('conv2'):
network = common.spatialConvolution(network, 1, 1, 64, wd=wd)
network = common.batchNormalization(network,
is_training=is_training)
network = tf.nn.relu(network)
with tf.variable_scope('conv3'):
network = common.spatialConvolution(network, 3, 1, 192, wd=wd)
network = common.batchNormalization(network,
is_training=is_training)
network = tf.nn.relu(network)
network = common.maxPool(network, 3, 2)
with tf.variable_scope('inception3a'):
network = inception(network,
64, [96, 128], [16, 32],
32,
wd=wd,
is_training=is_training)
with tf.variable_scope('inception3b'):
network = inception(network,
128, [128, 192], [32, 96],
64,
wd=wd,
is_training=is_training)
network = common.maxPool(network, 3, 2)
with tf.variable_scope('inception4a'):
network = inception(network,
192, [96, 208], [16, 48],
64,
wd=wd,
is_training=is_training)
with tf.variable_scope('inception4b'):
network = inception(network,
160, [112, 224], [24, 64],
64,
wd=wd,
is_training=is_training)
with tf.variable_scope('inception4c'):
network = inception(network,
128, [128, 256], [24, 64],
64,
wd=wd,
is_training=is_training)
with tf.variable_scope('inception4d'):
network = inception(network,
112, [144, 288], [32, 64],
64,
wd=wd,
is_training=is_training)
with tf.variable_scope('mainb'):
with tf.variable_scope('inception4e'):
main_branch = inception(network,
256, [160, 320], [32, 128],
128,
wd=wd,
is_training=is_training)
main_branch = common.maxPool(main_branch, 3, 2)
with tf.variable_scope('inception5a'):
main_branch = inception(main_branch,
256, [160, 320], [32, 128],
128,
wd=wd,
is_training=is_training)
with tf.variable_scope('inception5b'):
main_branch = inception(main_branch,
384, [192, 384], [48, 128],
128,
wd=wd,
is_training=is_training)
main_branch = common.avgPool(main_branch, 7, 1)
main_branch = common.flatten(main_branch)
main_branch = tf.nn.dropout(main_branch, dropout_rate)
if not transfer_mode:
with tf.variable_scope('output'):
main_branch = common.fullyConnected(main_branch,
num_output,
wd=wd)
else:
with tf.variable_scope('transfer_output'):
main_branch = common.fullyConnected(main_branch,
num_output,
wd=wd)
with tf.variable_scope('auxb'):
aux_classifier = common.avgPool(network, 5, 3)
with tf.variable_scope('conv1'):
aux_classifier = common.spatialConvolution(aux_classifier,
1,
1,
128,
wd=wd)
aux_classifier = common.batchNormalization(aux_classifier,
is_training=is_training)
aux_classifier = tf.nn.relu(aux_classifier)
aux_classifier = common.flatten(aux_classifier)
with tf.variable_scope('fc1'):
aux_classifier = common.fullyConnected(aux_classifier, 1024, wd=wd)
aux_classifier = tf.nn.dropout(aux_classifier, dropout_rate)
if not transfer_mode:
with tf.variable_scope('output'):
aux_classifier = common.fullyConnected(aux_classifier,
num_output,
wd=wd)
else:
with tf.variable_scope('transfer_output'):
aux_classifier = common.fullyConnected(aux_classifier,
num_output,
wd=wd)
return tf.concat([main_branch, aux_classifier], 1)