def setup_training(self, session):
"""
Apply sigmoid non-linearity to side layer ouputs + fuse layer outputs
Compute total loss := side_layer_loss + fuse_layer_loss
Compute predicted edge maps from fuse layer as pseudo performance metric to track
"""
self.predictions = []
# self.coord_pred = []
self.loss = 0
self.io.print_warning('Deep supervision application set to {}'.format(
self.cfgs['deep_supervision']))
for idx, b in enumerate(self.side_outputs):
output = tf.nn.sigmoid(b, name='output_{}'.format(idx))
#print("The {} layer's shape is:".format(idx), b.get_shape())
cost = sigmoid_cross_entropy_balanced(
b, self.edgemaps, name='cross_entropy{}'.format(idx))
self.predictions.append(output)
if self.cfgs['deep_supervision']:
self.loss += (self.cfgs['loss_weights'] * cost)
# self.coord_pred.append(self.fc3l)
fuse_output = tf.nn.sigmoid(self.fuse, name='fuse')
fuse_cost = sigmoid_cross_entropy_balanced(self.fuse,
self.edgemaps,
name='cross_entropy_fuse')
# coincide_cost = vertices_and_edges_loss(self.fuse, self.fc3l, self.coordinates)
coord_cost = tf.losses.huber_loss(self.coordinates, self.regressor)
self.predictions.append(fuse_output)
self.loss += (self.cfgs['loss_weights'] * fuse_cost)
self.loss += coord_cost
# self.loss += coincide_cost
pred = tf.cast(tf.greater(fuse_output, 0.5),
tf.int32,
name='predictions')
error = tf.cast(tf.not_equal(pred, tf.cast(self.edgemaps, tf.int32)),
tf.float32)
cd_error = tf.cast(tf.square(self.regressor - self.coordinates),
tf.float32)
self.cd_error = tf.reduce_mean(cd_error, name='regression_error')
self.error = tf.reduce_mean(error, name='pixel_error')
tf.summary.scalar('loss', self.loss)
tf.summary.scalar('error', self.error)
tf.summary.scalar('cd_error', self.cd_error)
self.merged_summary = tf.summary.merge_all()
self.train_writer = tf.summary.FileWriter(
self.cfgs['save_dir'] + '/train', session.graph)
self.val_writer = tf.summary.FileWriter(self.cfgs['save_dir'] + '/val')
def setup_training(self, session):
"""
Apply sigmoid non-linearity to side layer ouputs + fuse layer outputs
Compute total loss := side_layer_loss + fuse_layer_loss
Compute predicted edge maps from fuse layer as pseudo performance metric to track
"""
self.predictions = []
self.loss = 0
self.io.print_warning('Deep supervision application set to {}'.format(
self.cfgs['deep_supervision']))
for idx, b in enumerate(self.side_outputs):
#tf.summary.image('output_{}'.format(idx), b)
output = tf.nn.sigmoid(b, name='output_{}'.format(idx))
cost = sigmoid_cross_entropy_balanced(
b, self.edgemaps, name='cross_entropy{}'.format(idx))
self.predictions.append(output)
if self.cfgs['deep_supervision']:
self.loss += (self.cfgs['loss_weights'] * cost)
#tf.summary.image('fuse', self.fuse)
fuse_output = tf.nn.sigmoid(self.fuse, name='fuse')
tf.summary.image('fuseSigmoid', fuse_output)
fuse_cost = sigmoid_cross_entropy_balanced(self.fuse,
self.edgemaps,
name='cross_entropy_fuse')
#tf.summary.image('cvSource', self.images)
tf.summary.image('cvSourceEdge', self.edgemaps)
self.predictions.append(fuse_output)
self.loss += (self.cfgs['loss_weights'] * fuse_cost)
pred = tf.cast(tf.greater(fuse_output, 0.5),
tf.int32,
name='predictions')
error = tf.cast(tf.not_equal(pred, tf.cast(self.edgemaps, tf.int32)),
tf.float32)
self.error = tf.reduce_mean(error, name='pixel_error')
tf.summary.scalar('loss', self.loss)
tf.summary.scalar('error', self.error)
self.merged_summary = tf.summary.merge_all()
save_dir = self.saveDir
if not os.path.exists(save_dir):
os.makedirs(save_dir)
self.train_writer = tf.summary.FileWriter(save_dir + '/train',
session.graph)
self.val_writer = tf.summary.FileWriter(save_dir + '/val')
def setup_training(self, session):
"""
Apply sigmoid non-linearity to side layer ouputs + fuse layer outputs
Compute total loss := side_layer_loss + fuse_layer_loss
Compute predicted edge maps from fuse layer as pseudo performance metric to track
"""
deep_supervision=True
self.predictions = []
self.loss = 0
for idx, b in enumerate(self.side_outputs):
output = tf.nn.sigmoid(b, name='output_{}'.format(idx))
cost = sigmoid_cross_entropy_balanced(b, self.edgemaps, name='cross_entropy{}'.format(idx))
self.predictions.append(output)
if deep_supervision:
self.loss += (1.0 * cost)
fuse_output = tf.nn.sigmoid(self.fuse, name='fuse')
fuse_cost = sigmoid_cross_entropy_balanced(self.fuse, self.edgemaps, name='cross_entropy_fuse')
self.predictions.append(fuse_output)
self.loss += (1.0 * fuse_cost)
pred = tf.cast(tf.greater(fuse_output, 0.5), tf.int32, name='predictions')
error = tf.cast(tf.not_equal(pred, tf.cast(self.edgemaps, tf.int32)), tf.float32)
self.error = tf.reduce_mean(error, name='pixel_error')
tf.summary.scalar('loss', self.loss)
tf.summary.scalar('error', self.error)
self.merged_summary = tf.summary.merge_all()
self.train_writer = tf.summary.FileWriter(base_path + '/train', session.graph)
self.val_writer = tf.summary.FileWriter(base_path + '/val')