def model_fn(features, labels, mode, params):
x = tf.reshape(features, [-1, 99, 161, 1], name='input_incep4')
x_norm = tf.layers.batch_normalization(x, training=mode == tf.estimator.ModeKeys.TRAIN, name='x_norm')
if params['verbose_summary']:
tf.summary.image('input', x)
conv1 = tf.layers.conv2d(x_norm, filters=16, kernel_size=3, padding='same', activation=tf.nn.relu, name='conv1')
conv1b = tf.layers.conv2d(conv1, filters=16, kernel_size=3, activation=tf.nn.relu, name='conv1b')
pool1 = tf.layers.max_pooling2d(conv1b, pool_size=[2, 2], strides=2, name='pool1')
if params['verbose_summary']:
log_conv_kernel('conv1')
log_conv_kernel('conv1b')
tf.summary.image('pool1', pool1[:, :, :, 0:1])
incep2 = inception_block(conv1, name='incep2')
incep3 = inception_block(incep2, t1x1=4, t3x3=4, t5x5=4, tmp=4, name='incep3')
incep4 = inception_block(incep3, t1x1=8, t3x3=8, t5x5=8, tmp=8, name='incep4')
incep5 = inception_block(incep4, t1x1=16, t3x3=16, t5x5=16, tmp=16, name='incep5')
incep6 = inception_block(incep5, t1x1=20, t3x3=20, t5x5=20, tmp=20, name='incep6')
flat = flatten(incep6)
dropout4 = tf.layers.dropout(flat, rate=params['dropout_rate'], training=mode == tf.estimator.ModeKeys.TRAIN, name='dropout4')
dense4 = tf.layers.dense(dropout4, units=2048, activation=tf.nn.relu, name='dense4')
logits = tf.layers.dense(dense4, units=params['num_classes'], name='logits')
predictions = {
'classes': tf.argmax(logits, axis=1, name='prediction_classes'),
'probabilities': tf.nn.softmax(logits, name='prediction_softmax')
}
if mode == tf.estimator.ModeKeys.PREDICT:
return tf.estimator.EstimatorSpec(mode=mode, predictions={'predictions': predictions['probabilities']})
onehot_labels = tf.one_hot(indices=tf.cast(labels, tf.int32), depth=params['num_classes'], name='onehot_labels')
loss = tf.losses.softmax_cross_entropy(onehot_labels=onehot_labels, logits=logits)
tf.summary.scalar('loss', loss)
optimizer = tf.train.GradientDescentOptimizer(learning_rate=params['learning_rate'])
update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)
with tf.control_dependencies(update_ops):
train_op = optimizer.minimize(loss=loss, global_step=tf.train.get_global_step())
eval_metric_ops = {
'accuracy': tf.metrics.accuracy(labels=labels, predictions=predictions['classes'])
}
tf.summary.scalar('accuracy', eval_metric_ops['accuracy'][1])
return tf.estimator.EstimatorSpec(
mode=mode,
loss=loss,
train_op=train_op,
eval_metric_ops=eval_metric_ops
)
def conv_group(prev, filters, name='conv_group', verbose=False):
with tf.variable_scope(name):
conv = tf.layers.conv2d(prev,
filters=filters,
kernel_size=3,
padding='same',
activation=tf.nn.relu,
name='conv_same')
convb = tf.layers.conv2d(conv,
filters=filters,
kernel_size=3,
activation=tf.nn.relu,
name='convb')
convc = tf.layers.conv2d(convb,
filters=filters,
kernel_size=3,
activation=tf.nn.relu,
name='convc')
pool = tf.layers.max_pooling2d(convc,
pool_size=[2, 2],
strides=2,
name='pool')
if verbose:
log_conv_kernel('{}/conv_same'.format(name))
log_conv_kernel('{}/convb'.format(name))
log_conv_kernel('{}/convc'.format(name))
tf.summary.image('{}/pool'.format(name), pool[:, :, :, 0:1])
return pool
def model_fn(features, labels, mode, params):
x = tf.reshape(features, [-1, 125, 161, 2], name='input_incep100_lg')
x_norm = tf.layers.batch_normalization(x, training=mode == tf.estimator.ModeKeys.TRAIN, name='x_norm')
if params['verbose_summary']:
with tf.variable_scope('input'):
for sample in range(3):
tf.summary.image('input_spec_{}'.format(sample), tf.reshape(x[sample, :, :, 0], [-1, 125, 161, 1]))
tf.summary.image('input_freq_{}'.format(sample), tf.reshape(x[sample, :, :, 1], [-1, 125, 161, 1]))
conv1 = tf.layers.conv2d(x_norm, filters=16, kernel_size=3, padding='same', activation=tf.nn.relu, name='conv1')
conv1b = tf.layers.conv2d(conv1, filters=16, kernel_size=3, activation=tf.nn.relu, name='conv1b')
pool1 = tf.layers.max_pooling2d(conv1b, pool_size=[2, 2], strides=2, name='pool1')
if params['verbose_summary']:
log_conv_kernel('conv1')
log_conv_kernel('conv1b')
tf.summary.image('pool1', pool1[:, :, :, 0:1])
incep2 = inception_block(pool1, t1x1=8, t3x3=8, t5x5=8, tmp=8, name='incep2')
conv3 = tf.layers.conv2d(incep2, filters=32, kernel_size=3, padding='same', activation=tf.nn.relu, name='conv3')
conv3b = tf.layers.conv2d(conv3, filters=32, kernel_size=3, activation=tf.nn.relu, name='conv3b')
pool3 = tf.layers.max_pooling2d(conv3b, pool_size=[2, 2], strides=2, name='pool3')
if params['verbose_summary']:
log_conv_kernel('conv3')
log_conv_kernel('conv3b')
tf.summary.image('pool3', pool3[:, :, :, 0:1])
conv5 = tf.layers.conv2d(pool3, filters=64, kernel_size=3, padding='same', activation=tf.nn.relu, name='conv5')
conv5b = tf.layers.conv2d(conv5, filters=64, kernel_size=3, activation=tf.nn.relu, name='conv5b')
pool5 = tf.layers.max_pooling2d(conv5b, pool_size=[2, 2], strides=2, name='pool5')
if params['verbose_summary']:
log_conv_kernel('conv5')
log_conv_kernel('conv5b')
tf.summary.image('pool5', pool5[:, :, :, 0:1])
incep6 = inception_block(pool5, t1x1=32, t3x3=32, t5x5=32, tmp=32, name='incep6')
conv7 = tf.layers.conv2d(incep6, filters=128, kernel_size=3, padding='same', activation=tf.nn.relu, name='conv7')
conv7b = tf.layers.conv2d(conv7, filters=128, kernel_size=3, activation=tf.nn.relu, name='conv7b')
pool7 = tf.layers.max_pooling2d(conv7b, pool_size=[2, 2], strides=2, name='pool7')
if params['verbose_summary']:
log_conv_kernel('conv7')
log_conv_kernel('conv7b')
tf.summary.image('pool7', pool7[:, :, :, 0:1])
conv8 = tf.layers.conv2d(pool7, filters=256, kernel_size=3, padding='same', activation=tf.nn.relu, name='conv8')
conv8b = tf.layers.conv2d(conv8, filters=256, kernel_size=3, activation=tf.nn.relu, name='conv8b')
pool8 = tf.layers.max_pooling2d(conv8b, pool_size=[2, 2], strides=2, name='pool8')
if params['verbose_summary']:
log_conv_kernel('conv8')
log_conv_kernel('conv8b')
tf.summary.image('pool8', pool8[:, :, :, 0:1])
flat = flatten(pool8)
dense4 = tf.layers.dense(flat, units=2048, activation=tf.nn.relu, name='dense4')
dropout4 = tf.layers.dropout(dense4, rate=params['dropout_rate'], training=mode == tf.estimator.ModeKeys.TRAIN, name='dropout4')
logits = tf.layers.dense(dropout4, units=params['num_classes'], name='logits')
predictions = {
'classes': tf.argmax(logits, axis=1, name='prediction_classes'),
'probabilities': tf.nn.softmax(logits, name='prediction_softmax')
}
if mode == tf.estimator.ModeKeys.PREDICT:
return tf.estimator.EstimatorSpec(mode=mode, predictions={'predictions': predictions['probabilities']})
onehot_labels = tf.one_hot(indices=tf.cast(labels, tf.int32), depth=params['num_classes'], name='onehot_labels')
loss = tf.losses.softmax_cross_entropy(onehot_labels=onehot_labels, logits=logits)
tf.summary.scalar('loss', loss)
optimizer = tf.train.GradientDescentOptimizer(learning_rate=params['learning_rate'])
update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)
with tf.control_dependencies(update_ops):
train_op = optimizer.minimize(loss=loss, global_step=tf.train.get_global_step())
eval_metric_ops = {
'accuracy': tf.metrics.accuracy(labels=labels, predictions=predictions['classes'])
}
tf.summary.scalar('accuracy', eval_metric_ops['accuracy'][1])
return tf.estimator.EstimatorSpec(
mode=mode,
loss=loss,
train_op=train_op,
eval_metric_ops=eval_metric_ops
)
def branch_incep(x, name, mode, params):
with tf.variable_scope(name):
conv1 = tf.layers.conv2d(x,
filters=16,
kernel_size=3,
padding='same',
activation=tf.nn.relu,
name='conv1')
conv1b = tf.layers.conv2d(conv1,
filters=16,
kernel_size=3,
activation=tf.nn.relu,
name='conv1b')
pool1 = tf.layers.max_pooling2d(conv1b,
pool_size=[2, 2],
strides=2,
name='pool1')
if params['verbose_summary']:
log_conv_kernel('{}/conv1'.format(name))
log_conv_kernel('{}/conv1b'.format(name))
tf.summary.image('pool1', pool1[:, :, :, 0:1])
incep2 = inception_block(pool1,
t1x1=8,
t3x3=8,
t5x5=8,
tmp=8,
name='incep2')
conv3 = tf.layers.conv2d(incep2,
filters=32,
kernel_size=3,
padding='same',
activation=tf.nn.relu,
name='conv3')
conv3b = tf.layers.conv2d(conv3,
filters=32,
kernel_size=3,
activation=tf.nn.relu,
name='conv3b')
pool3 = tf.layers.max_pooling2d(conv3b,
pool_size=[2, 2],
strides=2,
name='pool3')
if params['verbose_summary']:
log_conv_kernel('{}/conv3'.format(name))
log_conv_kernel('{}/conv3b'.format(name))
tf.summary.image('pool3', pool3[:, :, :, 0:1])
conv5 = tf.layers.conv2d(pool3,
filters=64,
kernel_size=3,
padding='same',
activation=tf.nn.relu,
name='conv5')
conv5b = tf.layers.conv2d(conv5,
filters=64,
kernel_size=3,
activation=tf.nn.relu,
name='conv5b')
pool5 = tf.layers.max_pooling2d(conv5b,
pool_size=[2, 2],
strides=2,
name='pool5')
if params['verbose_summary']:
log_conv_kernel('{}/conv5'.format(name))
log_conv_kernel('{}/conv5b'.format(name))
tf.summary.image('pool5', pool5[:, :, :, 0:1])
incep6 = inception_block(pool5,
t1x1=32,
t3x3=32,
t5x5=32,
tmp=32,
name='incep6')
conv7 = tf.layers.conv2d(incep6,
filters=128,
kernel_size=3,
padding='same',
activation=tf.nn.relu,
name='conv7')
conv7b = tf.layers.conv2d(conv7,
filters=128,
kernel_size=3,
activation=tf.nn.relu,
name='conv7b')
pool7 = tf.layers.max_pooling2d(conv7b,
pool_size=[2, 2],
strides=2,
name='pool7')
if params['verbose_summary']:
log_conv_kernel('{}/conv7'.format(name))
log_conv_kernel('{}/conv7b'.format(name))
tf.summary.image('pool7', pool7[:, :, :, 0:1])
conv8 = tf.layers.conv2d(pool7,
filters=256,
kernel_size=3,
padding='same',
activation=tf.nn.relu,
name='conv8')
conv8b = tf.layers.conv2d(conv8,
filters=256,
kernel_size=3,
activation=tf.nn.relu,
name='conv8b')
pool8 = tf.layers.max_pooling2d(conv8b,
pool_size=[2, 2],
strides=2,
name='pool8')
if params['verbose_summary']:
log_conv_kernel('{}/conv8'.format(name))
log_conv_kernel('{}/conv8b'.format(name))
tf.summary.image('pool8', pool8[:, :, :, 0:1])
return pool8
def model_fn(features, labels, mode, params):
x = tf.reshape(features, [-1, 125, 161, 2], name='cnn4')
x_norm = tf.layers.batch_normalization(
x, training=mode == tf.estimator.ModeKeys.TRAIN, name='x_norm')
x = tf.reshape(x_norm[:, :, :, 0], [-1, 125, 161, 1], name='reshape_spec')
if params['verbose_summary']:
tf.summary.image('input', x)
conv1 = tf.layers.conv2d(x,
filters=16,
kernel_size=5,
activation=tf.nn.relu,
name='conv1')
pool1 = tf.layers.max_pooling2d(conv1,
pool_size=[2, 2],
strides=2,
name='pool1')
if params['verbose_summary']:
log_conv_kernel('conv1')
tf.summary.image('pool1', pool1[:, :, :, 0:1])
conv2 = tf.layers.conv2d(pool1,
filters=32,
kernel_size=5,
activation=tf.nn.relu,
name='conv2')
pool2 = tf.layers.max_pooling2d(conv2,
pool_size=[2, 2],
strides=2,
name='pool2')
if params['verbose_summary']:
log_conv_kernel('conv2')
tf.summary.image('pool2', pool2[:, :, :, 0:1])
conv3 = tf.layers.conv2d(pool2,
filters=64,
kernel_size=5,
activation=tf.nn.relu,
name='conv3')
pool3 = tf.layers.max_pooling2d(conv3,
pool_size=[2, 2],
strides=2,
name='pool3')
if params['verbose_summary']:
log_conv_kernel('conv3')
tf.summary.image('pool3', pool3[:, :, :, 0:1])
dim = pool3.get_shape()[1:]
dim = int(dim[0] * dim[1] * dim[2])
flat = tf.reshape(pool3, [-1, dim], name='flat')
dropout4 = tf.layers.dropout(flat,
rate=params['dropout_rate'],
training=mode == ModeKeys.TRAIN,
name='dropout4')
dense4 = tf.layers.dense(dropout4,
units=128,
activation=tf.nn.relu,
name='dense4')
logits = tf.layers.dense(dense4,
units=params['num_classes'],
name='logits')
predictions = {
'classes': tf.argmax(logits, axis=1, name='prediction_classes'),
'probabilities': tf.nn.softmax(logits,
name='prediction_probabilities'),
}
if mode == ModeKeys.PREDICT:
return EstimatorSpec(
mode=mode,
predictions={'predictions': predictions['probabilities']})
tf.summary.image(
'confusion_matrix',
conf_mat(labels, predictions['classes'], params['num_classes']))
onehot_labels = tf.one_hot(indices=tf.cast(labels, tf.int32),
depth=params['num_classes'],
name='onehot_labels')
loss = tf.losses.softmax_cross_entropy(onehot_labels=onehot_labels,
logits=logits)
tf.summary.scalar('loss', loss)
optimizer = tf.train.GradientDescentOptimizer(
learning_rate=params['learning_rate'])
update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)
with tf.control_dependencies(update_ops):
train_op = optimizer.minimize(loss=loss,
global_step=tf.train.get_global_step())
eval_metric_ops = {
'accuracy':
tf.metrics.accuracy(labels=labels, predictions=predictions['classes'])
}
tf.summary.scalar('accuracy', eval_metric_ops['accuracy'][1])
return EstimatorSpec(mode=mode,
loss=loss,
train_op=train_op,
eval_metric_ops=eval_metric_ops)
def model_fn(features, labels, mode, params):
x = tf.reshape(features, [-1, 125, 128, 1], name='input_flatv2')
x_flat = tf.reshape(features, [-1, 16000])
x_norm = tf.layers.batch_normalization(
x, training=mode == tf.estimator.ModeKeys.TRAIN, name='x_norm')
if params['verbose_summary']:
tf.summary.image('input', x)
tf.summary.audio('input', x_flat, 16000)
conv1 = tf.layers.conv2d(x_norm,
filters=16,
kernel_size=7,
activation=tf.nn.relu,
name='conv1')
conv2 = tf.layers.conv2d(conv1,
filters=32,
kernel_size=5,
activation=tf.nn.relu,
name='conv2')
conv3 = tf.layers.conv2d(conv2,
filters=64,
kernel_size=3,
activation=tf.nn.relu,
name='conv3')
pool3 = tf.layers.max_pooling2d(conv3,
pool_size=[2, 2],
strides=2,
name='pool3')
if params['verbose_summary']:
for i in range(1, 4):
label = 'conv{}'.format(i)
graph_utils.log_conv_kernel(label)
tf.summary.image(label, tf.expand_dims(conv1[..., 0], -1))
tf.summary.image('pool3', pool3[:, :, :, 0:1])
conv4 = tf.layers.conv2d(pool3,
filters=128,
kernel_size=3,
activation=tf.nn.relu,
name='conv4')
conv5 = tf.layers.conv2d(conv4,
filters=256,
kernel_size=3,
activation=tf.nn.relu,
name='conv5')
conv6 = tf.layers.conv2d(conv5,
filters=512,
kernel_size=3,
activation=tf.nn.relu,
name='conv6')
pool6 = tf.layers.max_pooling2d(conv6,
pool_size=[2, 2],
strides=2,
name='pool6')
if params['verbose_summary']:
for i in range(4, 7):
label = 'conv{}'.format(i)
graph_utils.log_conv_kernel(label)
tf.summary.image(label, tf.expand_dims(conv1[..., 0], -1))
tf.summary.image('pool6', pool6[:, :, :, 0:1])
conv7 = tf.layers.conv2d(pool6,
filters=1024,
kernel_size=3,
activation=tf.nn.relu,
name='conv7')
conv8 = tf.layers.conv2d(conv7,
filters=1024,
kernel_size=3,
activation=tf.nn.relu,
name='conv8')
conv9 = tf.layers.conv2d(conv8,
filters=1024,
kernel_size=3,
activation=tf.nn.relu,
name='conv9')
pool9 = tf.layers.max_pooling2d(conv9,
pool_size=[2, 2],
strides=2,
name='pool9')
if params['verbose_summary']:
for i in range(7, 10):
label = 'conv{}'.format(i)
graph_utils.log_conv_kernel(label)
tf.summary.image(label, tf.expand_dims(conv1[..., 0], -1))
tf.summary.image('pool9', pool9[:, :, :, 0:1])
conv10 = tf.layers.conv2d(pool9,
filters=512,
kernel_size=1,
activation=tf.nn.relu,
name='conv10')
conv11 = tf.layers.conv2d(conv10,
filters=512,
kernel_size=1,
activation=tf.nn.relu,
name='conv11')
conv12 = tf.layers.conv2d(conv11,
filters=512,
kernel_size=1,
activation=tf.nn.relu,
name='conv12')
flat = flatten(conv12)
dense = tf.layers.dense(flat,
units=1024,
activation=tf.nn.relu,
name='dense')
dropout = tf.layers.dropout(dense,
rate=params['dropout_rate'],
training=mode == tf.estimator.ModeKeys.TRAIN,
name='dropout')
logits = tf.layers.dense(dropout, units=12, name='logits')
predictions = {
'classes': tf.argmax(logits, axis=1, name='prediction_classes'),
'probabilities': tf.nn.softmax(logits, name='prediction_softmax')
}
if mode == tf.estimator.ModeKeys.PREDICT:
return tf.estimator.EstimatorSpec(
mode=mode,
predictions={'predictions': predictions['probabilities']})
onehot_labels = tf.one_hot(indices=tf.cast(labels, tf.int32),
depth=12,
name='onehot_labels')
loss = tf.losses.softmax_cross_entropy(onehot_labels=onehot_labels,
logits=logits)
tf.summary.scalar('loss', loss)
optimizer = tf.train.GradientDescentOptimizer(
learning_rate=params['learning_rate'])
update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)
with tf.control_dependencies(update_ops):
train_op = optimizer.minimize(loss=loss,
global_step=tf.train.get_global_step())
eval_metric_ops = {
'accuracy':
tf.metrics.accuracy(labels=labels, predictions=predictions['classes'])
}
tf.summary.scalar('accuracy', eval_metric_ops['accuracy'][1])
return tf.estimator.EstimatorSpec(mode=mode,
loss=loss,
train_op=train_op,
eval_metric_ops=eval_metric_ops)
def model_fn(features, labels, mode, params):
x = tf.reshape(features, [-1, 99, 161, 1], name='input_incep14')
x_norm = tf.layers.batch_normalization(
x, training=mode == tf.estimator.ModeKeys.TRAIN, name='x_norm')
if params['verbose_summary']:
tf.summary.image('input', x)
conv1 = tf.layers.conv2d(x_norm,
filters=16,
kernel_size=3,
activation=tf.nn.relu,
name='conv1')
conv2 = tf.layers.conv2d(conv1,
filters=32,
kernel_size=3,
activation=tf.nn.relu,
name='conv2')
conv3 = tf.layers.conv2d(conv2,
filters=32,
kernel_size=3,
strides=(2, 2),
activation=tf.nn.relu,
name='conv3')
if params['verbose_summary']:
log_conv_kernel('conv1')
log_conv_kernel('conv2')
log_conv_kernel('conv3')
tf.summary.image('conv3', conv3[:, :, :, 0:1])
conv4 = tf.layers.conv2d(conv3,
filters=64,
kernel_size=3,
activation=tf.nn.relu,
name='conv4')
conv5 = tf.layers.conv2d(conv4,
filters=128,
kernel_size=3,
activation=tf.nn.relu,
name='conv5')
conv6 = tf.layers.conv2d(conv5,
filters=128,
kernel_size=3,
strides=(2, 2),
activation=tf.nn.relu,
name='conv6')
if params['verbose_summary']:
log_conv_kernel('conv4')
log_conv_kernel('conv5')
log_conv_kernel('conv6')
tf.summary.image('conv6', conv6[:, :, :, 0:1])
incep4 = inception_block_v2(conv6,
t1x1=32,
t3x3=32,
t5x5=32,
tmp=32,
name='incep4',
norm=True,
mode=mode)
incep5 = inception_block_v2(incep4,
t1x1=48,
t3x3=48,
t5x5=48,
tmp=48,
name='incep5',
norm=True,
mode=mode)
incep6 = inception_block_v2(incep5,
t1x1=64,
t3x3=64,
t5x5=64,
tmp=64,
name='incep6',
norm=True,
mode=mode)
# incep7 = inception_block_v2(incep6, t1x1=96, t3x3=96, t5x5=96, tmp=96, name='incep7', norm=True, mode=mode)
incep8 = inception_block_v2(incep6,
t1x1=128,
t3x3=128,
t5x5=128,
tmp=128,
name='incep8',
norm=True,
mode=mode)
# incep9 = inception_block_v2(incep8, t1x1=192, t3x3=192, t5x5=192, tmp=192, name='incep9', norm=True, mode=mode)
incep10 = inception_block_v2(incep8,
t1x1=256,
t3x3=256,
t5x5=256,
tmp=256,
name='incep10',
norm=True,
mode=mode)
# incep11 = inception_block_v2(incep10, t1x1=512, t3x3=512, t5x5=512, tmp=512, name='incep11', norm=True, mode=mode)
# incep12 = inception_block_v2(incep11, t1x1=768, t3x3=768, t5x5=768, tmp=768, name='incep12', norm=True, mode=mode)
flat = flatten(incep10)
dense13 = tf.layers.dense(flat,
units=2048,
activation=tf.nn.relu,
name='dense13')
dropout13 = tf.layers.dropout(dense13,
rate=params['dropout_rate'],
training=mode == tf.estimator.ModeKeys.TRAIN,
name='dropout13')
logits = tf.layers.dense(dropout13,
units=params['num_classes'],
name='logits')
predictions = {
'classes': tf.argmax(logits, axis=1, name='prediction_classes'),
'probabilities': tf.nn.softmax(logits, name='prediction_softmax')
}
if mode == tf.estimator.ModeKeys.PREDICT:
return tf.estimator.EstimatorSpec(
mode=mode,
predictions={'predictions': predictions['probabilities']})
onehot_labels = tf.one_hot(indices=tf.cast(labels, tf.int32),
depth=params['num_classes'],
name='onehot_labels')
loss = tf.losses.softmax_cross_entropy(onehot_labels=onehot_labels,
logits=logits)
tf.summary.scalar('loss', loss)
optimizer = tf.train.GradientDescentOptimizer(
learning_rate=params['learning_rate'])
update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS)
with tf.control_dependencies(update_ops):
train_op = optimizer.minimize(loss=loss,
global_step=tf.train.get_global_step())
eval_metric_ops = {
'accuracy':
tf.metrics.accuracy(labels=labels, predictions=predictions['classes'])
}
tf.summary.scalar('accuracy', eval_metric_ops['accuracy'][1])
return tf.estimator.EstimatorSpec(mode=mode,
loss=loss,
train_op=train_op,
eval_metric_ops=eval_metric_ops)
def branch(x, name, mode, params):
with tf.variable_scope(name):
conv1 = tf.layers.conv2d(x,
filters=16,
kernel_size=3,
padding='same',
activation=tf.nn.relu,
name='conv1')
conv1b = tf.layers.conv2d(conv1,
filters=16,
kernel_size=3,
activation=tf.nn.relu,
name='conv1b')
pool1 = tf.layers.max_pooling2d(conv1b,
pool_size=[2, 2],
strides=2,
name='pool1')
if params['verbose_summary']:
log_conv_kernel('{}/conv1'.format(name))
log_conv_kernel('{}/conv1b'.format(name))
tf.summary.image('pool1', pool1[:, :, :, 0:1])
incep2 = inception_block(pool1,
t1x1=8,
t3x3=8,
t5x5=8,
tmp=8,
name='incep2')
conv3 = tf.layers.conv2d(incep2,
filters=32,
kernel_size=3,
padding='same',
activation=tf.nn.relu,
name='conv3')
conv3b = tf.layers.conv2d(conv3,
filters=32,
kernel_size=3,
activation=tf.nn.relu,
name='conv3b')
pool3 = tf.layers.max_pooling2d(conv3b,
pool_size=[2, 2],
strides=2,
name='pool3')
if params['verbose_summary']:
log_conv_kernel('{}/conv3'.format(name))
log_conv_kernel('{}/conv3b'.format(name))
tf.summary.image('pool3', pool3[:, :, :, 0:1])
conv5 = tf.layers.conv2d(pool3,
filters=64,
kernel_size=3,
padding='same',
activation=tf.nn.relu,
name='conv5')
conv5b = tf.layers.conv2d(conv5,
filters=64,
kernel_size=3,
activation=tf.nn.relu,
name='conv5b')
pool5 = tf.layers.max_pooling2d(conv5b,
pool_size=[2, 2],
strides=2,
name='pool5')
if params['verbose_summary']:
log_conv_kernel('{}/conv5'.format(name))
log_conv_kernel('{}/conv5b'.format(name))
tf.summary.image('pool5', pool5[:, :, :, 0:1])
incep6 = inception_block(pool5,
t1x1=32,
t3x3=32,
t5x5=32,
tmp=32,
name='incep6')
conv7 = tf.layers.conv2d(incep6,
filters=128,
kernel_size=3,
padding='same',
activation=tf.nn.relu,
name='conv7')
conv7b = tf.layers.conv2d(conv7,
filters=128,
kernel_size=3,
activation=tf.nn.relu,
name='conv7b')
pool7 = tf.layers.max_pooling2d(conv7b,
pool_size=[2, 2],
strides=2,
name='pool7')
if params['verbose_summary']:
log_conv_kernel('{}/conv7'.format(name))
log_conv_kernel('{}/conv7b'.format(name))
tf.summary.image('pool7', pool7[:, :, :, 0:1])
conv8 = tf.layers.conv2d(pool7,
filters=256,
kernel_size=3,
padding='same',
activation=tf.nn.relu,
name='conv8')
conv8b = tf.layers.conv2d(conv8,
filters=256,
kernel_size=3,
activation=tf.nn.relu,
name='conv8b')
pool8 = tf.layers.max_pooling2d(conv8b,
pool_size=[2, 2],
strides=2,
name='pool8')
if params['verbose_summary']:
log_conv_kernel('{}/conv8'.format(name))
log_conv_kernel('{}/conv8b'.format(name))
tf.summary.image('pool8', pool8[:, :, :, 0:1])
flat = flatten(pool8)
dense = tf.layers.dense(flat,
units=2048,
activation=tf.nn.relu,
name='dense')
dropout = tf.layers.dropout(
dense,
rate=params['dropout_rate'],
training=mode == tf.estimator.ModeKeys.TRAIN,
name='dropout')
return dropout
