def build_heatmap(dataset, heat_map):
"""Evaluate model on Dataset for a number of steps."""
with tf.Graph().as_default():
# Get images and labels from the dataset.
images, cords = image_processing.inputs(dataset, BATCH_SIZE)
# Number of classes in the Dataset label set plus 1.
# Label 0 is reserved for an (unused) background class.
num_classes = dataset.num_classes()
_, _, prob_ops = inception.inference(images, num_classes)
# Restore the moving average version of the learned variables for eval.
variable_averages = tf.train.ExponentialMovingAverage(
inception.MOVING_AVERAGE_DECAY)
variables_to_restore = variable_averages.variables_to_restore()
saver = tf.train.Saver(variables_to_restore)
# Build the summary operation based on the TF collection of Summaries.
summary_op = tf.summary.merge_all()
graph_def = tf.get_default_graph().as_graph_def()
summary_writer = tf.summary.FileWriter(FLAGS.eval_dir,
graph_def=graph_def)
# _eval_once(saver, summary_writer, accuracy, summary_op, confusion_matrix_op, logits, labels, dense_labels)
heat_map = generate_heatmap(saver, dataset, summary_writer, prob_ops,
cords, summary_op, heat_map)
return heat_map
def build_heatmap(dataset, heat_map, model_name, wsi_filename):
"""Evaluate model on Dataset for a number of steps."""
with tf.Graph().as_default():
# Get images and labels from the dataset.
images, cords = image_processing.inputs(dataset, BATCH_SIZE)
# Number of classes in the Dataset label set plus 1.
# Label 0 is reserved for an (unused) background class.
num_classes = dataset.num_classes()
_, _, prob_ops = inception.inference(images, num_classes)
# Restore the moving average version of the learned variables for eval.
variable_averages = tf.train.ExponentialMovingAverage(
inception.MOVING_AVERAGE_DECAY)
variables_to_restore = variable_averages.variables_to_restore()
saver = tf.train.Saver(variables_to_restore)
# Build the summary operation based on the TF collection of Summaries.
summary_op = tf.summary.merge_all()
graph_def = tf.get_default_graph().as_graph_def()
summary_writer = tf.summary.FileWriter(FLAGS.eval_dir, graph_def=graph_def)
# _eval_once(saver, summary_writer, accuracy, summary_op, confusion_matrix_op, logits, labels, dense_labels)
heat_map = generate_heatmap(saver, dataset, model_name, prob_ops, cords, heat_map, wsi_filename)
return heat_map
def _tower_loss(images, labels, num_classes, scope, reuse_variables=None):
"""Calculate the total loss on a single tower running the ImageNet model.
We perform 'batch splitting'. This means that we cut up a batch across
multiple GPU's. For instance, if the batch size = 32 and num_gpus = 2,
then each tower will operate on an batch of 16 images.
Args:
images: Images. 4D tensor of size [batch_size, FLAGS.image_size,
FLAGS.image_size, 3].
labels: 1-D integer Tensor of [batch_size].
num_classes: number of classes
scope: unique prefix string identifying the ImageNet tower, e.g.
'tower_0'.
Returns:
Tensor of shape [] containing the total loss for a batch of data
"""
# When fine-tuning a model, we do not restore the logits but instead we
# randomly initialize the logits. The number of classes in the output of the
# logit is the number of classes in specified Dataset.
restore_logits = not FLAGS.fine_tune
# Build inference Graph.
with tf.variable_scope(tf.get_variable_scope(), reuse=reuse_variables):
logits = inception.inference(images, num_classes, for_training=True,
restore_logits=restore_logits,
scope=scope)
# Build the portion of the Graph calculating the losses. Note that we will
# assemble the total_loss using a custom function below.
split_batch_size = images.get_shape().as_list()[0]
inception.loss(logits, labels, batch_size=split_batch_size)
# Assemble all of the losses for the current tower only.
losses = tf.get_collection(slim.losses.LOSSES_COLLECTION, scope)
# Calculate the total loss for the current tower.
regularization_losses = tf.get_collection(tf.GraphKeys.REGULARIZATION_LOSSES)
total_loss = tf.add_n(losses + regularization_losses, name='total_loss')
# Compute the moving average of all individual losses and the total loss.
loss_averages = tf.train.ExponentialMovingAverage(0.9, name='avg')
loss_averages_op = loss_averages.apply(losses + [total_loss])
# Attach a scalar summary to all individual losses and the total loss; do the
# same for the averaged version of the losses.
for l in losses + [total_loss]:
# Remove 'tower_[0-9]/' from the name in case this is a multi-GPU training
# session. This helps the clarity of presentation on TensorBoard.
loss_name = re.sub('%s_[0-9]*/' % inception.TOWER_NAME, '', l.op.name)
# Name each loss as '(raw)' and name the moving average version of the loss
# as the original loss name.
tf.summary.scalar(loss_name + ' (raw)', l)
tf.summary.scalar(loss_name, loss_averages.average(l))
with tf.control_dependencies([loss_averages_op]):
total_loss = tf.identity(total_loss)
return total_loss
def evaluate(dataset):
"""Evaluate model on Dataset for a number of steps."""
with tf.Graph().as_default():
# Get images and labels from the dataset.
images, labels = image_processing.inputs(dataset, BATCH_SIZE)
# Number of classes in the Dataset label set plus 1.
# Label 0 is reserved for an (unused) background class.
num_classes = dataset.num_classes()
# Build a Graph that computes the logits predictions from the
# inference model.
logits, _, _ = inception.inference(images, num_classes)
sparse_labels = tf.reshape(labels, [BATCH_SIZE, 1])
indices = tf.reshape(tf.range(BATCH_SIZE), [BATCH_SIZE, 1])
concated = tf.concat(1, [indices, sparse_labels])
num_classes = logits[0].get_shape()[-1].value
dense_labels = tf.sparse_to_dense(concated, [BATCH_SIZE, num_classes],
1, 0)
confusion_matrix_op = metrics.confusion_matrix(
labels, tf.argmax(logits, axis=1))
# false_positive_op = metrics.streaming_false_positives(logits, dense_labels)
# false_negative_op = metrics.streaming_false_negatives(logits, dense_labels)
# Calculate predictions.
accuracy = tf.nn.in_top_k(logits, labels, 1)
# Restore the moving average version of the learned variables for eval.
variable_averages = tf.train.ExponentialMovingAverage(
inception.MOVING_AVERAGE_DECAY)
variables_to_restore = variable_averages.variables_to_restore()
saver = tf.train.Saver(variables_to_restore)
# Build the summary operation based on the TF collection of Summaries.
summary_op = tf.summary.merge_all()
graph_def = tf.get_default_graph().as_graph_def()
summary_writer = tf.summary.FileWriter(FLAGS.eval_dir,
graph_def=graph_def)
while True:
# _eval_once(saver, summary_writer, accuracy, summary_op, confusion_matrix_op, logits, labels, dense_labels)
_eval_once(saver, summary_writer, accuracy, summary_op,
confusion_matrix_op)
if FLAGS.run_once:
break
time.sleep(FLAGS.eval_interval_secs)
def evaluate(dataset):
"""Evaluate model on Dataset for a number of steps."""
with tf.Graph().as_default():
# Get images and labels from the dataset.
images, labels = image_processing.inputs(dataset, BATCH_SIZE)
# Number of classes in the Dataset label set plus 1.
# Label 0 is reserved for an (unused) background class.
num_classes = dataset.num_classes()
# Build a Graph that computes the logits predictions from the
# inference model.
logits, _, _ = inception.inference(images, num_classes)
sparse_labels = tf.reshape(labels, [BATCH_SIZE, 1])
indices = tf.reshape(tf.range(BATCH_SIZE), [BATCH_SIZE, 1])
concated = tf.concat(1, [indices, sparse_labels])
num_classes = logits[0].get_shape()[-1].value
dense_labels = tf.sparse_to_dense(concated,
[BATCH_SIZE, num_classes],
1, 0)
confusion_matrix_op = metrics.confusion_matrix(labels, tf.argmax(logits, axis=1))
# false_positive_op = metrics.streaming_false_positives(logits, dense_labels)
# false_negative_op = metrics.streaming_false_negatives(logits, dense_labels)
# Calculate predictions.
accuracy = tf.nn.in_top_k(logits, labels, 1)
# Restore the moving average version of the learned variables for eval.
variable_averages = tf.train.ExponentialMovingAverage(
inception.MOVING_AVERAGE_DECAY)
variables_to_restore = variable_averages.variables_to_restore()
saver = tf.train.Saver(variables_to_restore)
# Build the summary operation based on the TF collection of Summaries.
summary_op = tf.summary.merge_all()
graph_def = tf.get_default_graph().as_graph_def()
summary_writer = tf.summary.FileWriter(FLAGS.eval_dir, graph_def=graph_def)
while True:
# _eval_once(saver, summary_writer, accuracy, summary_op, confusion_matrix_op, logits, labels, dense_labels)
_eval_once(saver, summary_writer, accuracy, summary_op, confusion_matrix_op)
if FLAGS.run_once:
break
time.sleep(FLAGS.eval_interval_secs)
def build_heatmap(dataset):
"""Evaluate model on Dataset for a number of steps."""
with tf.Graph().as_default():
# Get images and labels from the dataset.
images, cords = image_processing.inputs(dataset, BATCH_SIZE)
# Number of classes in the Dataset label set plus 1.
# Label 0 is reserved for an (unused) background class.
num_classes = dataset.num_classes()
assert BATCH_SIZE % FLAGS.num_threads == 0, 'BATCH_SIZE must be divisible by FLAGS.num_threads'
# Build a Graph that computes the logits predictions from the
# inference model.
images_splits = tf.split(images, FLAGS.num_threads, axis=0)
cords_splits = tf.split(cords, FLAGS.num_threads, axis=0)
prob_ops = []
cords_ops = []
for i in range(FLAGS.num_threads):
with tf.name_scope('%s_%d' % (inception.TOWER_NAME, i)) as scope:
with slim.arg_scope([slim.variables.variable],
device='/cpu:%d' % i):
print('i=%d' % i)
_, _, prob_op = inception.inference(images_splits[i],
num_classes,
scope=scope)
cords_op = tf.reshape(
cords_splits[i],
(int(BATCH_SIZE / FLAGS.num_threads), 1))
prob_ops.append(prob_op)
cords_ops.append(cords_op)
# Restore the moving average version of the learned variables for eval.
variable_averages = tf.train.ExponentialMovingAverage(
inception.MOVING_AVERAGE_DECAY)
variables_to_restore = variable_averages.variables_to_restore()
saver = tf.train.Saver(variables_to_restore)
# Build the summary operation based on the TF collection of Summaries.
summary_op = tf.summary.merge_all()
graph_def = tf.get_default_graph().as_graph_def()
summary_writer = tf.summary.FileWriter(FLAGS.eval_dir,
graph_def=graph_def)
generate_heatmap(saver, dataset, summary_writer, prob_ops, cords_ops,
summary_op)
def build_heatmap(dataset):
"""Evaluate model on Dataset for a number of steps."""
with tf.Graph().as_default():
# Get images and labels from the dataset.
images, cords = image_processing.inputs(dataset, BATCH_SIZE)
# Number of classes in the Dataset label set plus 1.
# Label 0 is reserved for an (unused) background class.
num_classes = dataset.num_classes()
assert BATCH_SIZE % FLAGS.num_threads == 0, 'BATCH_SIZE must be divisible by FLAGS.num_threads'
# Build a Graph that computes the logits predictions from the
# inference model.
images_splits = tf.split(images, FLAGS.num_threads, axis=0)
cords_splits = tf.split(cords, FLAGS.num_threads, axis=0)
prob_ops = []
cords_ops = []
for i in range(FLAGS.num_threads):
with tf.name_scope('%s_%d' % (inception.TOWER_NAME, i)) as scope:
with slim.arg_scope([slim.variables.variable], device='/cpu:%d' % i):
print('i=%d' % i)
_, _, prob_op = inception.inference(images_splits[i], num_classes, scope=scope)
cords_op = tf.reshape(cords_splits[i], (int(BATCH_SIZE/FLAGS.num_threads), 1))
prob_ops.append(prob_op)
cords_ops.append(cords_op)
# Restore the moving average version of the learned variables for eval.
variable_averages = tf.train.ExponentialMovingAverage(
inception.MOVING_AVERAGE_DECAY)
variables_to_restore = variable_averages.variables_to_restore()
saver = tf.train.Saver(variables_to_restore)
# Build the summary operation based on the TF collection of Summaries.
summary_op = tf.summary.merge_all()
graph_def = tf.get_default_graph().as_graph_def()
summary_writer = tf.summary.FileWriter(FLAGS.eval_dir, graph_def=graph_def)
generate_heatmap(saver, dataset, summary_writer, prob_ops, cords_ops, summary_op)
def evaluate(dataset):
"""Evaluate model on Dataset for a number of steps."""
with tf.Graph().as_default():
# Get images and labels from the dataset.
images, labels = image_processing.inputs(dataset, BATCH_SIZE)
# Number of classes in the Dataset label set plus 1.
# Label 0 is reserved for an (unused) background class.
num_classes = dataset.num_classes()
# Build a Graph that computes the logits predictions from the
# inference model.
logits, _ = inception.inference(images, num_classes)
print(logits.get_shape())
print(labels.get_shape())
# Calculate predictions.
top_1_op = tf.nn.in_top_k(logits, labels, 1)
top_5_op = tf.nn.in_top_k(logits, labels, 5)
# Restore the moving average version of the learned variables for eval.
variable_averages = tf.train.ExponentialMovingAverage(
inception.MOVING_AVERAGE_DECAY)
variables_to_restore = variable_averages.variables_to_restore()
saver = tf.train.Saver(variables_to_restore)
# Build the summary operation based on the TF collection of Summaries.
summary_op = tf.summary.merge_all()
graph_def = tf.get_default_graph().as_graph_def()
summary_writer = tf.summary.FileWriter(FLAGS.eval_dir,
graph_def=graph_def)
while True:
_eval_once(saver, summary_writer, top_1_op, top_5_op, summary_op)
if FLAGS.run_once:
break
time.sleep(FLAGS.eval_interval_secs)
def evaluate(dataset):
"""Evaluate model on Dataset for a number of steps."""
with tf.Graph().as_default():
# Get images and labels from the dataset.
images, labels = image_processing.inputs(dataset, BATCH_SIZE)
# Number of classes in the Dataset label set plus 1.
# Label 0 is reserved for an (unused) background class.
num_classes = dataset.num_classes()
# Build a Graph that computes the logits predictions from the
# inference model.
logits, _ = inception.inference(images, num_classes)
print(logits.get_shape())
print(labels.get_shape())
# Calculate predictions.
top_1_op = tf.nn.in_top_k(logits, labels, 1)
top_5_op = tf.nn.in_top_k(logits, labels, 5)
# Restore the moving average version of the learned variables for eval.
variable_averages = tf.train.ExponentialMovingAverage(
inception.MOVING_AVERAGE_DECAY)
variables_to_restore = variable_averages.variables_to_restore()
saver = tf.train.Saver(variables_to_restore)
# Build the summary operation based on the TF collection of Summaries.
summary_op = tf.summary.merge_all()
graph_def = tf.get_default_graph().as_graph_def()
summary_writer = tf.summary.FileWriter(FLAGS.eval_dir, graph_def=graph_def)
while True:
_eval_once(saver, summary_writer, top_1_op, top_5_op, summary_op)
if FLAGS.run_once:
break
time.sleep(FLAGS.eval_interval_secs)
def _tower_loss(images, labels, num_classes, scope, reuse_variables=None):
"""Calculate the total loss on a single tower running the ImageNet model.
We perform 'batch splitting'. This means that we cut up a batch across
multiple GPU's. For instance, if the batch size = 32 and num_gpus = 2,
then each tower will operate on an batch of 16 images.
Args:
images: Images. 4D tensor of size [batch_size, FLAGS.image_size,
FLAGS.image_size, 3].
labels: 1-D integer Tensor of [batch_size].
num_classes: number of classes
scope: unique prefix string identifying the ImageNet tower, e.g.
'tower_0'.
Returns:
Tensor of shape [] containing the total loss for a batch of data
"""
# When fine-tuning a model, we do not restore the logits but instead we
# randomly initialize the logits. The number of classes in the output of the
# logit is the number of classes in specified Dataset.
restore_logits = not FLAGS.fine_tune
# Build inference Graph.
with tf.variable_scope(tf.get_variable_scope(), reuse=reuse_variables):
logits = inception.inference(images,
num_classes,
for_training=True,
restore_logits=restore_logits,
scope=scope)
# Build the portion of the Graph calculating the losses. Note that we will
# assemble the total_loss using a custom function below.
split_batch_size = images.get_shape().as_list()[0]
inception.loss(logits, labels, batch_size=split_batch_size)
# Assemble all of the losses for the current tower only.
losses = tf.get_collection(slim.losses.LOSSES_COLLECTION, scope)
# Calculate the total loss for the current tower.
regularization_losses = tf.get_collection(
tf.GraphKeys.REGULARIZATION_LOSSES)
total_loss = tf.add_n(losses + regularization_losses, name='total_loss')
# Compute the moving average of all individual losses and the total loss.
loss_averages = tf.train.ExponentialMovingAverage(0.9, name='avg')
loss_averages_op = loss_averages.apply(losses + [total_loss])
# Attach a scalar summary to all individual losses and the total loss; do the
# same for the averaged version of the losses.
for l in losses + [total_loss]:
# Remove 'tower_[0-9]/' from the name in case this is a multi-GPU training
# session. This helps the clarity of presentation on TensorBoard.
loss_name = re.sub('%s_[0-9]*/' % inception.TOWER_NAME, '', l.op.name)
# Name each loss as '(raw)' and name the moving average version of the loss
# as the original loss name.
tf.summary.scalar(loss_name + ' (raw)', l)
tf.summary.scalar(loss_name, loss_averages.average(l))
with tf.control_dependencies([loss_averages_op]):
total_loss = tf.identity(total_loss)
return total_loss
