def encode_coordinates_fn(self, net):
"""
Adds one-hot encoding of coordinates to different views in the networks.
For each "pixel" of a feature map it adds a one hot encoded x and y
coordinates.
:param net: a tensor of shape=[batch_size, height, width, num_features]
:return: a tensor with the same height and width, but altered feature_size.
"""
mparams = self._mparams['encode_coordinates_fn']
if mparams.enabled:
batch_size, h, w, _ = net.shape.as_list()
# create two matrix has shape (w, h) or (w, h)
x, y = tf.meshgrid(tf.range(w), tf.range(h))
w_loc = tf_slim.one_hot_encoding(
x, num_classes=w) # shape of (w, h, w)
h_loc = tf_slim.one_hot_encoding(
y, num_classes=h) # shape of (w, h, h)
loc = tf.concat([h_loc, w_loc], axis=2) # shape of (w, h, w + h)
loc = tf.tile(
tf.expand_dims(loc, 0),
[batch_size, 1, 1, 1]) # shape of (batch_size, w, h, w + h)
return tf.concat(
[net, loc],
3) # shape of (batch_size, w, h, w + h + num_features)
else:
return net
def char_prediction(self, chars_logit):
"""
return confidence scores (softmax values) for predicted characters
:param chars_logit: chars logits, a tensor with shape [batch_size x seq_length x num_char_classes]
:return:
A tuple (ids, log_prob, scores), where:
ids - predicted characters, a int32 tensor with shape
[batch_size x seq_length];
log_prob - a log probability of all characters, a float tensor with
shape [batch_size, seq_length, num_char_classes];
scores - corresponding confidence scores for characters, a float
tensor with shape [batch_size x seq_length].
"""
log_prob = logits_to_log_prob(chars_logit)
ids = tf.cast(tf.argmax(log_prob, axis=2),
name='predicted_chars',
dtype=tf.int32)
mask = tf.cast(
tf_slim.one_hot_encoding(ids, self._params.num_char_classes),
tf.bool)
all_scores = tf.nn.softmax(chars_logit)
selected_scores = tf.boolean_mask(all_scores, mask, name='char_scores')
scores = tf.reshape(selected_scores,
shape=(-1, self._params.seq_length))
return ids, log_prob, scores
def imagenet_input(is_training):
"""Data reader for imagenet.
Reads in imagenet data and performs pre-processing on the images.
Args:
is_training: bool specifying if train or validation dataset is needed.
Returns:
A batch of images and labels.
"""
if is_training:
dataset = dataset_factory.get_dataset('imagenet', 'train',
FLAGS.dataset_dir)
else:
dataset = dataset_factory.get_dataset('imagenet', 'validation',
FLAGS.dataset_dir)
provider = slim.dataset_data_provider.DatasetDataProvider(
dataset,
shuffle=is_training,
common_queue_capacity=2 * FLAGS.batch_size,
common_queue_min=FLAGS.batch_size)
[image, label] = provider.get(['image', 'label'])
image_preprocessing_fn = preprocessing_factory.get_preprocessing(
'mobilenet_v1', is_training=is_training)
image = image_preprocessing_fn(image, FLAGS.image_size, FLAGS.image_size)
images, labels = tf.train.batch([image, label],
batch_size=FLAGS.batch_size,
num_threads=4,
capacity=5 * FLAGS.batch_size)
labels = slim.one_hot_encoding(labels, FLAGS.num_classes)
return images, labels
def get_data(dataset,
batch_size,
augment=False,
central_crop_size=None,
shuffle_config=None,
shuffle=True):
"""Wraps calls to DatasetDataProviders and shuffle_batch.
For more details about supported Dataset objects refer to datasets/fsns.py.
Args:
dataset: a slim.data.dataset.Dataset object.
batch_size: number of samples per batch.
augment: optional, if True does random image distortion.
central_crop_size: A CharLogit tuple (crop_width, crop_height).
shuffle_config: A namedtuple ShuffleBatchConfig.
shuffle: if True use data shuffling.
Returns:
"""
if not shuffle_config:
shuffle_config = DEFAULT_SHUFFLE_CONFIG
provider = tf_slim.dataset_data_provider.DatasetDataProvider(
dataset,
shuffle=shuffle,
common_queue_capacity=2 * batch_size,
common_queue_min=batch_size)
image_orig, label = provider.get(['image', 'label'])
image = preprocess_image(image_orig,
augment,
central_crop_size,
num_towers=dataset.num_of_views)
label_one_hot = tf_slim.one_hot_encoding(label, dataset.num_char_classes)
# print(image.get_shape())
# print(image_orig.get_shape())
# print(label[0].get_shape())
# print(label_one_hot.get_shape())
"""
dataset = tf.data.Dataset.from_tensor_slices((image, image_orig, label, label_one_hot))
dataset = dataset.shuffle(buffer_size=shuffle_config.min_after_dequeue, reshuffle_each_iteration=True).batch(batch_size=batch_size)
images = tf.constant(list(dataset.map(lambda x_img, x_img_orig, y_label, y_label_one_hot: x_img)))
images_orig = tf.constant(list(dataset.map(lambda x_img, x_img_orig, y_label, y_label_one_hot: x_img_orig)))
labels = tf.constant(list(dataset.map(lambda x_img, x_img_orig, y_label, y_label_one_hot: y_label)))
labels_one_hot = tf.constant(list(dataset.map(lambda x_img, x_img_orig, y_label, y_label_one_hot: y_label_one_hot)))
"""
images, images_orig, labels, labels_one_hot = (
tf.compat.v1.train.shuffle_batch(
[image, image_orig, label, label_one_hot],
batch_size=batch_size,
num_threads=shuffle_config.num_batching_threads,
capacity=shuffle_config.queue_capacity,
min_after_dequeue=shuffle_config.min_after_dequeue))
return InputEndpoints(images=images,
images_orig=images_orig,
labels=labels,
labels_one_hot=labels_one_hot)
def char_one_hot(self, logit):
"""Creates one hot encoding for a logit of a character.
Args:
logit: A tensor with shape [batch_size, num_char_classes].
Returns:
A tensor with shape [batch_size, num_char_classes]
"""
prediction = tf.argmax(logit, axis=1)
return tf_slim.one_hot_encoding(prediction, self._params.num_char_classes)
def build_model(self):
tf.reset_default_graph()
self.losses = []
self.vars = []
self.avg_gradient = []
self.apply_grad = []
self.instances = []
self.gradients = []
class setter():
def __init__(self, assignment, devices):
self.assignment = assignment
self.last_device = devices[0]
def choose(self, op):
scope = tf.get_variable_scope().name
for key in self.assignment:
if key in scope:
self.last_device = self.assignment[key]
return self.assignment[key]
#print(self.assignment)
print(scope, op.name, self.last_device)
return self.last_device
def device_setter(assignment, devices):
_setter = setter(assignment, devices)
return _setter.choose
losses = []
outputs = []
tf.get_variable_scope()._reuse = tf.AUTO_REUSE
for i in range(1):
loss, output, scopes = self.model_fn(None, self.model_name)
losses.append(loss)
outputs.append(output[-1])
self.scopes = scopes
new_loss = tf.add_n(losses)
new_loss = tf.reduce_mean(new_loss, name="final_loss")
#self.train_op = tf.train.AdamOptimizer(learning_rate=0.2, beta1=0.9, beta2=0.98, epsilon=1e-9).minimize(new_loss)
self.train_op = tf.train.GradientDescentOptimizer(
learning_rate=0.01).minimize(new_loss,
colocate_gradients_with_ops=True)
init = tf.global_variables_initializer()
g = tf.get_default_graph().as_graph_def(add_shapes=True)
import tge
strategy = {node.name: [1, 1, 1, 1, 1] for node in g.node}
g = (
tge.TGE(g, devices).custom(strategy)
# .replace_placeholder(BATCHSIZE)
.use_collective()
# .verbose()
.compile().get_result())
with open("vgg_tge_modified.pbtxt", "w") as fo:
fo.write(pbtf.MessageToString(g))
tf.reset_default_graph()
gdef = graph_pb2.GraphDef()
with open("vgg_tge_modified.pbtxt", "r") as f:
txt = f.read()
pbtf.Parse(txt, gdef)
tf.import_graph_def(gdef)
graph = tf.get_default_graph()
dataset = dataset_factory.get_dataset("imagenet", "train",
"/data/slim_imagenet")
preprocessing_name = "vgg_19"
image_preprocessing_fn = preprocessing_factory.get_preprocessing(
preprocessing_name, is_training=True)
provider = slim.dataset_data_provider.DatasetDataProvider(
dataset,
num_readers=4,
common_queue_capacity=20 * batch_size,
common_queue_min=10 * batch_size)
[image, label] = provider.get(['image', 'label'])
train_image_size = 224
image = image_preprocessing_fn(image, train_image_size,
train_image_size)
print("image shape:", image.shape)
print("label shape:", label.shape)
images, labels = tf.train.batch([image, label],
batch_size=batch_size,
num_threads=4,
capacity=5 * batch_size)
labels = slim.one_hot_encoding(labels, dataset.num_classes)
batch_queue = slim.prefetch_queue.prefetch_queue([images, labels],
capacity=2 *
micro_batch_num)
x_tensor = graph.get_tensor_by_name("import/Placeholder/replica_0:0")
y_tensor = graph.get_tensor_by_name("import/Placeholder_1/replica_0:0")
x, y = batch_queue.dequeue()
replace_input(graph, x, x_tensor.name)
replace_input(graph, y, y_tensor.name)
opt = graph.get_operation_by_name("import/GradientDescent/replica_0")
loss = tf.reduce_mean(tf.add_n(get_tensors(graph, "final_loss")))
init = graph.get_operation_by_name("import/init/replica_0")
config = tf.ConfigProto()
config.allow_soft_placement = True
sess = tf.Session(config=config)
sess.run(init)
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
for i in range(10000000):
_, cal_loss = sess.run([opt, loss])
if i % 10 == 0:
print("Step:{},Loss:{}".format(i, cal_loss))
def build_model(self):
tf.reset_default_graph()
self.losses = []
self.vars = []
self.avg_gradient = []
self.apply_grad = []
self.instances = []
self.gradients = []
gpu_num = 4
recorded_accuracy5 = []
global_start_time = time.time()
with open("vgg_dp3_time_record.txt", "w") as f:
f.write("global start time: {}\n".format(global_start_time))
times = []
class setter():
def __init__(self, assignment, devices):
self.assignment = assignment
self.last_device = devices[0]
def choose(self, op):
scope = tf.get_variable_scope().name
for key in self.assignment:
if key in scope:
self.last_device = self.assignment[key]
return self.assignment[key]
#print(self.assignment)
print(scope, op.name, self.last_device)
return self.last_device
def device_setter(assignment, devices):
_setter = setter(assignment, devices)
return _setter.choose
losses = []
outputs = []
with tf.variable_scope("input", reuse=tf.AUTO_REUSE):
dataset = dataset_factory.get_dataset("imagenet", "train",
"/data/slim_imagenet")
preprocessing_name = "vgg_19"
image_preprocessing_fn = preprocessing_factory.get_preprocessing(
preprocessing_name, is_training=True)
provider = slim.dataset_data_provider.DatasetDataProvider(
dataset,
num_readers=4,
common_queue_capacity=20 * batch_size,
common_queue_min=10 * batch_size)
[image, label] = provider.get(['image', 'label'])
train_image_size = 224
image = image_preprocessing_fn(image, train_image_size,
train_image_size)
print("image shape:", image.shape)
print("label shape:", label.shape)
images, labels = tf.train.batch([image, label],
batch_size=batch_size,
num_threads=4,
capacity=5 * batch_size)
labels = slim.one_hot_encoding(labels, dataset.num_classes)
batch_queue = slim.prefetch_queue.prefetch_queue([images, labels],
capacity=2 *
gpu_num)
tf.get_variable_scope()._reuse = tf.AUTO_REUSE
for i in range(gpu_num):
with tf.device("gpu:{}".format(i)):
loss, output, scopes = self.model_fn(batch_queue,
self.model_name)
losses.append(loss)
outputs.append(output[-1])
self.scopes = scopes
with tf.device("gpu:2"):
new_loss = tf.add_n(losses, name="final_loss") / gpu_num
new_loss = tf.reduce_mean(new_loss)
new_outputs = tf.add_n(outputs)
#self.train_op = tf.train.AdamOptimizer(learning_rate=0.2, beta1=0.9, beta2=0.98, epsilon=1e-9).minimize(new_loss)
#self.train_op = tf.train.GradientDescentOptimizer(learning_rate=0.01).minimize(new_loss,colocate_gradients_with_ops=True)
self.train_op = tf.train.MomentumOptimizer(
learning_rate=0.01,
momentum=0.9).minimize(new_loss, colocate_gradients_with_ops=True)
graph = tf.get_default_graph()
accurate_num = get_tensors(graph, "top_accuracy")
print("accurate_num:", accurate_num)
#accurate_num = tf.reduce_sum(tf.add_n(accurate_num))
accurate_num = tf.reduce_sum(accurate_num[0])
init = tf.global_variables_initializer()
config = tf.ConfigProto()
config.allow_soft_placement = True
sess = tf.Session(config=config)
sess.run(init)
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
start_time = time.time()
for i in range(10000000):
_, loss, accuracy_num = sess.run(
[self.train_op, new_loss, accurate_num])
#top5accuracy = accuracy_num / (gpu_num * batch_size)
top5accuracy = accuracy_num / (batch_size)
if i % 10 == 0:
end_time = time.time()
print(
"Step:{},Loss:{},top5 accuracy:{},per_step_time:{}".format(
i, loss, top5accuracy, (end_time - start_time) / 10))
start_time = time.time()
gap = top5accuracy * 100 // 5 * 5
if gap not in recorded_accuracy5:
global_end_time = time.time()
recorded_accuracy5.append(gap)
print(
"achieveing {}% at the first time, concreate top5 accuracy: {}%. time slot: {}, duration: {}s\n"
.format(gap, top5accuracy * 100, global_end_time,
global_end_time - global_start_time),
flush=True)
with open("vgg_dp3_time_record.txt", "a+") as f:
f.write(
"achieveing {}% at the first time, concreate top5 accuracy: {}%. time slot: {}, duration: {}s\n"
.format(gap, top5accuracy * 100, global_end_time,
global_end_time - global_start_time))
def main(_):
if not FLAGS.dataset_dir:
raise ValueError(
'You must supply the dataset directory with --dataset_dir')
tf.logging.set_verbosity(tf.logging.INFO)
with tf.Graph().as_default():
#######################
# Config model_deploy #
#######################
deploy_config = model_deploy.DeploymentConfig(
num_clones=FLAGS.num_clones,
clone_on_cpu=FLAGS.clone_on_cpu,
replica_id=FLAGS.task,
num_replicas=FLAGS.worker_replicas,
num_ps_tasks=FLAGS.num_ps_tasks)
# Create global_step
with tf.device(deploy_config.variables_device()):
global_step = slim.create_global_step()
######################
# Select the dataset #
######################
dataset = dataset_factory.get_dataset(FLAGS.dataset_name,
FLAGS.dataset_split_name,
FLAGS.dataset_dir)
######################
# Select the network #
######################
network_fn = nets_factory.get_network_fn(
FLAGS.model_name,
num_classes=(dataset.num_classes - FLAGS.labels_offset),
weight_decay=FLAGS.weight_decay,
is_training=True)
#####################################
# Select the preprocessing function #
#####################################
preprocessing_name = FLAGS.preprocessing_name or FLAGS.model_name
image_preprocessing_fn = preprocessing_factory.get_preprocessing(
preprocessing_name, is_training=True)
##############################################################
# Create a dataset provider that loads data from the dataset #
##############################################################
with tf.device(deploy_config.inputs_device()):
provider = slim.dataset_data_provider.DatasetDataProvider(
dataset,
num_readers=FLAGS.num_readers,
common_queue_capacity=20 * FLAGS.batch_size,
common_queue_min=10 * FLAGS.batch_size)
[image, label] = provider.get(['image', 'label'])
label -= FLAGS.labels_offset
train_image_size = FLAGS.train_image_size or network_fn.default_image_size
image = image_preprocessing_fn(image, train_image_size,
train_image_size)
images, labels = tf.train.batch(
[image, label],
batch_size=FLAGS.batch_size,
num_threads=FLAGS.num_preprocessing_threads,
capacity=5 * FLAGS.batch_size)
labels = slim.one_hot_encoding(
labels, dataset.num_classes - FLAGS.labels_offset)
batch_queue = slim.prefetch_queue.prefetch_queue(
[images, labels], capacity=2 * deploy_config.num_clones)
####################
# Define the model #
####################
def clone_fn(batch_queue):
"""Allows data parallelism by creating multiple clones of network_fn."""
images, labels = batch_queue.dequeue()
logits, end_points = network_fn(images)
#############################
# Specify the loss function #
#############################
if 'AuxLogits' in end_points:
slim.losses.softmax_cross_entropy(
end_points['AuxLogits'],
labels,
label_smoothing=FLAGS.label_smoothing,
weights=0.4,
scope='aux_loss')
slim.losses.softmax_cross_entropy(
logits,
labels,
label_smoothing=FLAGS.label_smoothing,
weights=1.0)
return end_points
# Gather initial summaries.
summaries = set(tf.get_collection(tf.compat.v1.GraphKeys.SUMMARIES))
clones = model_deploy.create_clones(deploy_config, clone_fn,
[batch_queue])
first_clone_scope = deploy_config.clone_scope(0)
# Gather update_ops from the first clone. These contain, for example,
# the updates for the batch_norm variables created by network_fn.
update_ops = tf.get_collection(tf.compat.v1.GraphKeys.UPDATE_OPS,
first_clone_scope)
# Add summaries for end_points.
end_points = clones[0].outputs
for end_point in end_points:
x = end_points[end_point]
summaries.add(tf.summary.histogram('activations/' + end_point, x))
summaries.add(
tf.summary.scalar('sparsity/' + end_point,
tf.nn.zero_fraction(x)))
# Add summaries for losses.
for loss in tf.get_collection(tf.compat.v1.GraphKeys.LOSSES,
first_clone_scope):
summaries.add(tf.summary.scalar('losses/%s' % loss.op.name, loss))
# Add summaries for variables.
for variable in slim.get_model_variables():
summaries.add(tf.summary.histogram(variable.op.name, variable))
#################################
# Configure the moving averages #
#################################
if FLAGS.moving_average_decay:
moving_average_variables = slim.get_model_variables()
variable_averages = tf.train.ExponentialMovingAverage(
FLAGS.moving_average_decay, global_step)
else:
moving_average_variables, variable_averages = None, None
#if FLAGS.quantize_delay >= 0:
# tf.contrib.quantize.create_training_graph(
# quant_delay=FLAGS.quantize_delay)
#########################################
# Configure the optimization procedure. #
#########################################
with tf.device(deploy_config.optimizer_device()):
learning_rate = _configure_learning_rate(dataset.num_samples,
global_step)
optimizer = _configure_optimizer(learning_rate)
summaries.add(tf.summary.scalar('learning_rate', learning_rate))
if FLAGS.sync_replicas:
# If sync_replicas is enabled, the averaging will be done in the chief
# queue runner.
optimizer = tf.train.SyncReplicasOptimizer(
opt=optimizer,
replicas_to_aggregate=FLAGS.replicas_to_aggregate,
total_num_replicas=FLAGS.worker_replicas,
variable_averages=variable_averages,
variables_to_average=moving_average_variables)
elif FLAGS.moving_average_decay:
# Update ops executed locally by trainer.
update_ops.append(
variable_averages.apply(moving_average_variables))
# Variables to train.
variables_to_train = _get_variables_to_train()
# and returns a train_tensor and summary_op
total_loss, clones_gradients = model_deploy.optimize_clones(
clones, optimizer, var_list=variables_to_train)
# Add total_loss to summary.
summaries.add(tf.summary.scalar('total_loss', total_loss))
# Create gradient updates.
grad_updates = optimizer.apply_gradients(clones_gradients,
global_step=global_step)
update_ops.append(grad_updates)
update_op = tf.group(*update_ops)
with tf.control_dependencies([update_op]):
train_tensor = tf.identity(total_loss, name='train_op')
# Add the summaries from the first clone. These contain the summaries
# created by model_fn and either optimize_clones() or _gather_clone_loss().
summaries |= set(
tf.get_collection(tf.compat.v1.GraphKeys.SUMMARIES,
first_clone_scope))
# Merge all summaries together.
summary_op = tf.summary.merge(list(summaries), name='summary_op')
###########################
# Kicks off the training. #
###########################
slim.learning.train(
train_tensor,
logdir=FLAGS.train_dir,
master=FLAGS.master,
is_chief=(FLAGS.task == 0),
init_fn=_get_init_fn(),
summary_op=summary_op,
number_of_steps=FLAGS.max_number_of_steps,
log_every_n_steps=FLAGS.log_every_n_steps,
save_summaries_secs=FLAGS.save_summaries_secs,
save_interval_secs=FLAGS.save_interval_secs,
sync_optimizer=optimizer if FLAGS.sync_replicas else None)
def main(_):
#tf.disable_v2_behavior() ###
tf.compat.v1.disable_eager_execution()
tf.compat.v1.enable_resource_variables()
# Enable habana bf16 conversion pass
if FLAGS.dtype == 'bf16':
os.environ['TF_BF16_CONVERSION'] = flags.FLAGS.bf16_config_path
FLAGS.precision = 'bf16'
else:
os.environ['TF_BF16_CONVERSION'] = "0"
if FLAGS.use_horovod:
hvd_init()
if not FLAGS.dataset_dir:
raise ValueError(
'You must supply the dataset directory with --dataset_dir')
tf.logging.set_verbosity(tf.logging.INFO)
with tf.Graph().as_default():
#######################
# Config model_deploy #
#######################
deploy_config = model_deploy.DeploymentConfig(
num_clones=FLAGS.num_clones,
clone_on_cpu=FLAGS.clone_on_cpu,
replica_id=FLAGS.task,
num_replicas=FLAGS.worker_replicas,
num_ps_tasks=FLAGS.num_ps_tasks)
# Create global_step
with tf.device(deploy_config.variables_device()):
global_step = slim.create_global_step()
######################
# Select the dataset #
######################
dataset = dataset_factory.get_dataset(FLAGS.dataset_name,
FLAGS.dataset_split_name,
FLAGS.dataset_dir)
######################
# Select the network #
######################
network_fn = nets_factory.get_network_fn(
FLAGS.model_name,
num_classes=(dataset.num_classes - FLAGS.labels_offset),
weight_decay=FLAGS.weight_decay,
is_training=True)
#####################################
# Select the preprocessing function #
#####################################
preprocessing_name = FLAGS.preprocessing_name or FLAGS.model_name
image_preprocessing_fn = preprocessing_factory.get_preprocessing(
preprocessing_name,
is_training=True,
use_grayscale=FLAGS.use_grayscale)
##############################################################
# Create a dataset provider that loads data from the dataset #
##############################################################
with tf.device(deploy_config.inputs_device()):
provider = slim.dataset_data_provider.DatasetDataProvider(
dataset,
num_readers=FLAGS.num_readers,
common_queue_capacity=20 * FLAGS.batch_size,
common_queue_min=10 * FLAGS.batch_size)
[image, label] = provider.get(['image', 'label'])
label -= FLAGS.labels_offset
train_image_size = FLAGS.train_image_size or network_fn.default_image_size
image = image_preprocessing_fn(image, train_image_size,
train_image_size)
images, labels = tf.train.batch(
[image, label],
batch_size=FLAGS.batch_size,
num_threads=FLAGS.num_preprocessing_threads,
capacity=5 * FLAGS.batch_size)
labels = slim.one_hot_encoding(
labels, dataset.num_classes - FLAGS.labels_offset)
batch_queue = slim.prefetch_queue.prefetch_queue(
[images, labels], capacity=2 * deploy_config.num_clones)
####################
# Define the model #
####################
def clone_fn(batch_queue):
"""Allows data parallelism by creating multiple clones of network_fn."""
images, labels = batch_queue.dequeue()
logits, end_points = network_fn(images)
#############################
# Specify the loss function #
#############################
if 'AuxLogits' in end_points:
slim.losses.softmax_cross_entropy(
end_points['AuxLogits'],
labels,
label_smoothing=FLAGS.label_smoothing,
weights=0.4,
scope='aux_loss')
slim.losses.softmax_cross_entropy(
logits,
labels,
label_smoothing=FLAGS.label_smoothing,
weights=1.0)
return end_points
# Gather initial summaries.
summaries = set(tf.get_collection(tf.GraphKeys.SUMMARIES))
clones = model_deploy.create_clones(deploy_config, clone_fn,
[batch_queue])
first_clone_scope = deploy_config.clone_scope(0)
# Gather update_ops from the first clone. These contain, for example,
# the updates for the batch_norm variables created by network_fn.
update_ops = tf.get_collection(tf.GraphKeys.UPDATE_OPS,
first_clone_scope)
# Add summaries for end_points.
end_points = clones[0].outputs
for end_point in end_points:
x = end_points[end_point]
summaries.add(tf.summary.histogram('activations/' + end_point, x))
summaries.add(
tf.summary.scalar('sparsity/' + end_point,
tf.nn.zero_fraction(x)))
# Add summaries for variables.
for variable in slim.get_model_variables():
summaries.add(tf.summary.histogram(variable.op.name, variable))
#################################
# Configure the moving averages #
#################################
if FLAGS.moving_average_decay:
moving_average_variables = slim.get_model_variables()
variable_averages = tf.train.ExponentialMovingAverage(
FLAGS.moving_average_decay, global_step)
else:
moving_average_variables, variable_averages = None, None
#if FLAGS.quantize_delay >= 0:
# quantize.create_training_graph(quant_delay=FLAGS.quantize_delay) #for debugging!!
#########################################
# Configure the optimization procedure. #
#########################################
with tf.device(deploy_config.optimizer_device()):
learning_rate = _configure_learning_rate(dataset.num_samples,
global_step)
optimizer = _configure_optimizer(learning_rate)
summaries.add(tf.summary.scalar('learning_rate', learning_rate))
if FLAGS.sync_replicas:
# If sync_replicas is enabled, the averaging will be done in the chief
# queue runner.
optimizer = tf.train.SyncReplicasOptimizer(
opt=optimizer,
replicas_to_aggregate=FLAGS.replicas_to_aggregate,
total_num_replicas=FLAGS.worker_replicas,
variable_averages=variable_averages,
variables_to_average=moving_average_variables)
elif FLAGS.moving_average_decay:
# Update ops executed locally by trainer.
update_ops.append(
variable_averages.apply(moving_average_variables))
# Variables to train.
variables_to_train = _get_variables_to_train()
# and returns a train_tensor and summary_op
total_loss, clones_gradients = model_deploy.optimize_clones(
clones, optimizer, var_list=variables_to_train)
# Create gradient updates.
grad_updates = optimizer.apply_gradients(clones_gradients,
global_step=global_step)
update_ops.append(grad_updates)
update_op = tf.group(*update_ops)
with tf.control_dependencies([update_op]):
train_tensor = tf.identity(total_loss, name='train_op')
# Add the summaries from the first clone. These contain the summaries
# created by model_fn and either optimize_clones() or _gather_clone_loss().
summaries |= set(
tf.get_collection(tf.GraphKeys.SUMMARIES, first_clone_scope))
# Merge all summaries together.
summary_op = tf.summary.merge(list(summaries), name='summary_op')
if horovod_enabled():
hvd.broadcast_global_variables(0)
###########################
# Kicks off the training. #
###########################
with dump_callback():
with logger.benchmark_context(FLAGS):
eps1 = ExamplesPerSecondKerasHook(FLAGS.log_every_n_steps,
output_dir=FLAGS.train_dir,
batch_size=FLAGS.batch_size)
write_hparams_v1(
eps1.writer, {
'batch_size': FLAGS.batch_size,
**{x: getattr(FLAGS, x)
for x in FLAGS}
})
train_step_kwargs = {}
if FLAGS.max_number_of_steps:
should_stop_op = math_ops.greater_equal(
global_step, FLAGS.max_number_of_steps)
else:
should_stop_op = constant_op.constant(False)
train_step_kwargs['should_stop'] = should_stop_op
if FLAGS.log_every_n_steps > 0:
train_step_kwargs['should_log'] = math_ops.equal(
math_ops.mod(global_step, FLAGS.log_every_n_steps), 0)
eps1.on_train_begin()
train_step_kwargs['EPS'] = eps1
slim.learning.train(
train_tensor,
logdir=FLAGS.train_dir,
train_step_fn=train_step1,
train_step_kwargs=train_step_kwargs,
master=FLAGS.master,
is_chief=(FLAGS.task == 0),
init_fn=_get_init_fn(),
summary_op=summary_op,
summary_writer=None,
number_of_steps=FLAGS.max_number_of_steps,
log_every_n_steps=FLAGS.log_every_n_steps,
save_summaries_secs=FLAGS.save_summaries_secs,
save_interval_secs=FLAGS.save_interval_secs,
sync_optimizer=optimizer if FLAGS.sync_replicas else None)
def main(_):
tf.disable_eager_execution()
if not FLAGS.dataset_dir:
raise ValueError('You must supply the dataset directory with --dataset_dir')
tf.logging.set_verbosity(tf.logging.INFO)
with tf.Graph().as_default():
######################
# Select the dataset #
######################
dataset = dataset_factory.get_dataset(
FLAGS.dataset_name, FLAGS.dataset_split_name, FLAGS.dataset_dir)
#########################
# Configure the network #
#########################
inception_params = network_params.InceptionV3FCNParams(
receptive_field_size=FLAGS.receptive_field_size,
prelogit_dropout_keep_prob=0.8,
depth_multiplier=0.1,
min_depth=16,
inception_fcn_stride=0,
)
conv_params = network_params.ConvScopeParams(
dropout=False,
dropout_keep_prob=0.8,
batch_norm=True,
batch_norm_decay=0.99,
l2_weight_decay=4e-05,
)
network_fn = inception_v3_fcn.get_inception_v3_fcn_network_fn(
inception_params,
conv_params,
num_classes=dataset.num_classes,
is_training=True,
)
#####################################
# Select the preprocessing function #
#####################################
image_preprocessing_fn = preprocessing_factory.get_preprocessing(
'inception_v3', is_training=True)
##############################################################
# Create a dataset provider that loads data from the dataset #
##############################################################
provider = slim.dataset_data_provider.DatasetDataProvider(
dataset,
num_readers=DATASET_READERS,
common_queue_capacity=20 * FLAGS.batch_size,
common_queue_min=10 * FLAGS.batch_size)
[image, label] = provider.get(['image', 'label'])
train_image_size = FLAGS.receptive_field_size
image = image_preprocessing_fn(image, train_image_size, train_image_size)
images, labels = tf.train.batch([image, label],
batch_size=FLAGS.batch_size,
num_threads=PREPROCESSING_THREADS,
capacity=5 * FLAGS.batch_size)
labels = slim.one_hot_encoding(labels, dataset.num_classes)
####################
# Define the model #
####################
logits, _ = network_fn(images)
slim.losses.softmax_cross_entropy(logits, labels)
total_loss = slim.losses.get_total_loss()
tf.summary.scalar('losses/Total_Loss', total_loss)
optimizer = tf.train.RMSPropOptimizer(0.01)
train_op = slim.learning.create_train_op(
total_loss,
optimizer,
variables_to_train=_get_variables_to_train())
###########################
# Kicks off the training. #
###########################
slim.learning.train(
train_op,
logdir=FLAGS.train_dir,
init_fn=_get_init_fn(),
number_of_steps=FLAGS.max_number_of_steps,
log_every_n_steps=FLAGS.log_every_n_steps,
save_summaries_secs=FLAGS.save_summaries_secs,
save_interval_secs=FLAGS.save_interval_secs,
session_config=tf.ConfigProto(allow_soft_placement=True))
def activate_unit(self,path,graph_def):
#setup_workers(workers, "grpc+verbs")
tf.reset_default_graph()
#server = tf.distribute.Server(cluster, job_name='worker', task_index=0, protocol="grpc+verbs",
# config=config)
target = None
tf.import_graph_def(graph_def)
print("import success")
graph = tf.get_default_graph()
init0 = graph.get_operation_by_name("import/init/replica_0")
print("11111111111111111111111")
dataset = dataset_factory.get_dataset(
"imagenet", "train", "/data/slim_imagenet")
preprocessing_name = "vgg_19"
image_preprocessing_fn = preprocessing_factory.get_preprocessing(
preprocessing_name,
is_training=True)
provider = slim.dataset_data_provider.DatasetDataProvider(
dataset,
num_readers=4,
common_queue_capacity=20 * batch_size*micro_batch_num,
common_queue_min=10 * batch_size*micro_batch_num,)
[image, label] = provider.get(['image', 'label'])
train_image_size = 224
image = image_preprocessing_fn(image, train_image_size, train_image_size)
print("image shape:", image.shape)
print("label shape:", label.shape)
images, labels = tf.train.batch(
[image, label],
batch_size=batch_size*micro_batch_num,
num_threads=4,
capacity=5 * batch_size*micro_batch_num)
labels = slim.one_hot_encoding(
labels, dataset.num_classes)
batch_queue = slim.prefetch_queue.prefetch_queue(
[images, labels], capacity=2 * micro_batch_num)
input_dict = None
'''
placeholders = [node.outputs[0] for node in graph.get_operations() if node.node_def.op == 'Placeholder']
shapes = [(p.shape.as_list()) for p in placeholders ]
for shape in shapes:
shape[0]=batch_size
input_dict = { p: np.random.rand(*shapes[i]) for i,p in enumerate(placeholders) }
'''
#prepare input
xs = ["import/input/Placeholder/replica_0:0"]
ys = ["import/input/Placeholder_1/replica_0:0"]
for i in range(1,micro_batch_num):
xs.append("import/input_{}/Placeholder/replica_0:0".format(i))
ys.append("import/input_{}/Placeholder_1/replica_0:0".format(i))
x, y = batch_queue.dequeue()
for i in range(len(xs)):
replace_input(graph,x[i*batch_size:(i+1)*batch_size],xs[i])
replace_input(graph,y[i*batch_size:(i+1)*batch_size],ys[i])
losses = get_tensors(graph, "final_loss")
losses = tf.reduce_mean(tf.add_n(losses)/len(losses))
accurate_num = get_tensors(graph,"top_accuracy")
print("accurate_num:",accurate_num)
total_batch_size = batch_size*micro_batch_num
size_for_each = total_batch_size/len(accurate_num)
num_to_calculate = int(64/size_for_each)
accurate_num = tf.reduce_sum(tf.add_n(accurate_num[:num_to_calculate]))
config = tf.ConfigProto()
config.allow_soft_placement = True
sess = tf.Session(target, config=config) # , config=tf.ConfigProto(allow_soft_placement=False))
print("222222222222222222222222")
print("333333333333333333333")
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
opt = []
for sink in self.sinks:
op = graph.get_operation_by_name('import/' + sink + "/replica_0")
opt.append(op)
# opt = [graph.get_operation_by_name('import/' + x) for x in self.sinks]
print("444444444444444444444")
recorded_accuracy5 = []
global_start_time = time.time()
with open("time_record.txt", "w") as f:
f.write("global start time: {}\n".format(global_start_time))
times= []
sess.run(init0)
#sess.run(init1)
start_time = time.time()
for j in range(100000000000000):
ret = sess.run(opt + [losses,accurate_num], feed_dict=input_dict)
loss = ret[-2]
top5accuracy_num = ret[-1]
top5accuracy = top5accuracy_num/64
if j % 10 == 0:
end_time = time.time()
print("Step:{},Loss:{},top5 accuracy:{},per_step_time:{}".format(j,loss,top5accuracy,(end_time-start_time)/10))
start_time = time.time()
gap = top5accuracy*100 // 5 * 5
if gap not in recorded_accuracy5:
global_end_time = time.time()
recorded_accuracy5.append(gap)
print("achieveing {}% at the first time, concreate top5 accuracy: {}%. time slot: {}, duration: {}s\n".format(gap,top5accuracy*100,global_end_time,global_end_time-global_start_time),flush=True)
with open("time_record.txt","a+") as f:
f.write("achieveing {}% at the first time, concreate top5 accuracy: {}%. time slot: {}, duration: {}s\n".format(gap,top5accuracy*100,global_end_time,global_end_time-global_start_time))
avg_time = sum(times)/len(times)
print(path,times,"average time:", avg_time)
print(" ")
'''
def build_model(self):
tf.reset_default_graph()
self.losses = []
self.vars = []
self.avg_gradient = []
self.apply_grad = []
self.instances = []
self.gradients = []
class setter():
def __init__(self, assignment, devices):
self.assignment = assignment
self.last_device = devices[0]
def choose(self, op):
scope = tf.get_variable_scope().name
for key in self.assignment:
if key in scope:
self.last_device = self.assignment[key]
return self.assignment[key]
#print(self.assignment)
print(scope, op.name, self.last_device)
return self.last_device
def device_setter(assignment, devices):
_setter = setter(assignment, devices)
return _setter.choose
losses = []
outputs = []
with tf.variable_scope("input", reuse=tf.AUTO_REUSE):
dataset = dataset_factory.get_dataset("imagenet", "train",
"/data/slim_imagenet")
preprocessing_name = "vgg_19"
image_preprocessing_fn = preprocessing_factory.get_preprocessing(
preprocessing_name, is_training=True)
provider = slim.dataset_data_provider.DatasetDataProvider(
dataset,
num_readers=4,
common_queue_capacity=20 * batch_size,
common_queue_min=10 * batch_size)
[image, label] = provider.get(['image', 'label'])
train_image_size = 224
image = image_preprocessing_fn(image, train_image_size,
train_image_size)
print("image shape:", image.shape)
print("label shape:", label.shape)
images, labels = tf.train.batch([image, label],
batch_size=batch_size,
num_threads=4,
capacity=5 * batch_size)
labels = slim.one_hot_encoding(labels, dataset.num_classes)
batch_queue = slim.prefetch_queue.prefetch_queue([images, labels],
capacity=2 *
micro_batch_num)
tf.get_variable_scope()._reuse = tf.AUTO_REUSE
for i in range(1):
with tf.device("gpu:{}".format(i)):
loss, output, scopes = self.model_fn(batch_queue,
self.model_name)
losses.append(loss)
outputs.append(output[-1])
self.scopes = scopes
with tf.device("gpu:0"):
new_loss = tf.add_n(losses, name="final_loss")
new_loss = tf.reduce_mean(new_loss)
new_outputs = tf.add_n(outputs)
#self.train_op = tf.train.AdamOptimizer(learning_rate=0.2, beta1=0.9, beta2=0.98, epsilon=1e-9).minimize(new_loss)
self.train_op = tf.train.GradientDescentOptimizer(
learning_rate=0.01).minimize(new_loss,
colocate_gradients_with_ops=True)
init = tf.global_variables_initializer()
config = tf.ConfigProto()
config.allow_soft_placement = True
sess = tf.Session(config=config)
sess.run(init)
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
for i in range(10000000):
_, loss = sess.run([self.train_op, new_loss])
if i % 10 == 0:
print("Step:{},Loss:{}".format(i, loss))
def main(model_root, datasets_dir, model_name):
# tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.INFO)
# 训练相关参数设置
with tf.Graph().as_default():
deploy_config = model_deploy.DeploymentConfig(
num_clones=num_clones,
clone_on_cpu=False,
replica_id=task,
num_replicas=worker_replicas,
num_ps_tasks=num_ps_tasks)
global_step = slim.create_global_step()
train_dir = os.path.join(model_root, model_name)
dataset = convert_data.get_datasets('train', dataset_dir=datasets_dir)
network_fn = net_select.get_network_fn(model_name,
num_classes=dataset.num_classes,
weight_decay=weight_decay,
is_training=True)
image_preprocessing_fn = preprocessing_select.get_preprocessing(
model_name, is_training=True)
print("the data_sources:", dataset.data_sources)
with tf.device(deploy_config.inputs_device()):
provider = slim.dataset_data_provider.DatasetDataProvider(
dataset,
num_readers=num_readers,
common_queue_capacity=20 * batch_size,
common_queue_min=10 * batch_size)
[image, label] = provider.get(['image', 'label'])
train_image_size = network_fn.default_image_size
image = image_preprocessing_fn(image, train_image_size,
train_image_size)
images, labels = tf.compat.v1.train.batch(
[image, label],
batch_size=batch_size,
num_threads=num_preprocessing_threads,
capacity=5 * batch_size)
labels = slim.one_hot_encoding(labels, dataset.num_classes)
batch_queue = slim.prefetch_queue.prefetch_queue(
[images, labels], capacity=2 * deploy_config.num_clones)
def calculate_pooling_center_loss(features, label, alfa, nrof_classes,
weights, name):
features = tf.reshape(features, [features.shape[0], -1])
label = tf.argmax(label, 1)
nrof_features = features.get_shape()[1]
centers = tf.compat.v1.get_variable(
name, [nrof_classes, nrof_features],
dtype=tf.float32,
initializer=tf.constant_initializer(0),
trainable=False)
label = tf.reshape(label, [-1])
centers_batch = tf.gather(centers, label)
centers_batch = tf.nn.l2_normalize(centers_batch, axis=-1)
diff = (1 - alfa) * (centers_batch - features)
centers = tf.compat.v1.scatter_sub(centers, label, diff)
with tf.control_dependencies([centers]):
distance = tf.square(features - centers_batch)
distance = tf.reduce_sum(distance, axis=-1)
center_loss = tf.reduce_mean(distance)
center_loss = tf.identity(center_loss * weights,
name=name + '_loss')
return center_loss
def attention_crop(attention_maps):
'''
利用attention map 做数据增强,这里是论文中的Crop Mask
:param attention_maps: Feature maps降维得到的
:return:
'''
batch_size, height, width, num_parts = attention_maps.shape
bboxes = []
for i in range(batch_size):
attention_map = attention_maps[i]
part_weights = attention_map.mean(axis=0).mean(axis=0)
part_weights = np.sqrt(part_weights)
part_weights = part_weights / np.sum(part_weights)
selected_index = np.random.choice(np.arange(0, num_parts),
1,
p=part_weights)[0]
mask = attention_map[:, :, selected_index]
threshold = random.uniform(0.4, 0.6)
itemindex = np.where(mask >= mask.max() * threshold)
ymin = itemindex[0].min() / height - 0.1
ymax = itemindex[0].max() / height + 0.1
xmin = itemindex[1].min() / width - 0.1
xmax = itemindex[1].max() / width + 0.1
bbox = np.asarray([ymin, xmin, ymax, xmax], dtype=np.float32)
bboxes.append(bbox)
bboxes = np.asarray(bboxes, np.float32)
return bboxes
def attention_drop(attention_maps):
'''
这里是attention drop部分，目的是为了让模型可以注意到物体的其他部位（因不同attention map可能聚焦了同一部位）
:param attention_maps:
:return:
'''
batch_size, height, width, num_parts = attention_maps.shape
masks = []
for i in range(batch_size):
attention_map = attention_maps[i]
part_weights = attention_map.mean(axis=0).mean(axis=0)
part_weights = np.sqrt(part_weights)
if (np.sum(part_weights) != 0):
part_weights = part_weights / np.sum(part_weights)
selected_index = np.random.choice(np.arange(0, num_parts),
1,
p=part_weights)[0]
mask = attention_map[:, :, selected_index:selected_index + 1]
# soft mask
threshold = random.uniform(0.2, 0.5)
mask = (mask < threshold * mask.max()).astype(np.float32)
masks.append(mask)
masks = np.asarray(masks, dtype=np.float32)
return masks
def clone_fn(batch_queue):
"""Allows data parallelism by creating multiple clones of network_fn."""
images, labels = batch_queue.dequeue()
logits_1, end_points_1 = network_fn(images)
attention_maps = end_points_1['attention_maps']
attention_maps = tf.image.resize(
attention_maps, [train_image_size, train_image_size],
method=tf.image.ResizeMethod.BILINEAR)
# attention crop
bboxes = tf.compat.v1.py_func(attention_crop, [attention_maps],
[tf.float32])
bboxes = tf.reshape(bboxes, [batch_size, 4])
box_ind = tf.range(batch_size, dtype=tf.int32)
images_crop = tf.image.crop_and_resize(
images,
bboxes,
box_ind,
crop_size=[train_image_size, train_image_size])
# attention drop
masks = tf.compat.v1.py_func(attention_drop, [attention_maps],
[tf.float32])
masks = tf.reshape(
masks, [batch_size, train_image_size, train_image_size, 1])
images_drop = images * masks
logits_2, end_points_2 = network_fn(images_crop, reuse=True)
logits_3, end_points_3 = network_fn(images_drop, reuse=True)
slim.losses.softmax_cross_entropy(logits_1,
labels,
weights=1 / 3.0,
scope='cross_entropy_1')
slim.losses.softmax_cross_entropy(logits_2,
labels,
weights=1 / 3.0,
scope='cross_entropy_2')
slim.losses.softmax_cross_entropy(logits_3,
labels,
weights=1 / 3.0,
scope='cross_entropy_3')
embeddings = end_points_1['embeddings']
center_loss = calculate_pooling_center_loss(
features=embeddings,
label=labels,
alfa=0.95,
nrof_classes=dataset.num_classes,
weights=1.0,
name='center_loss')
slim.losses.add_loss(center_loss)
return end_points_1
# Gather initial summaries.
summaries = set(
tf.compat.v1.get_collection(tf.compat.v1.GraphKeys.SUMMARIES))
clones = model_deploy.create_clones(deploy_config, clone_fn,
[batch_queue])
first_clone_scope = deploy_config.clone_scope(0)
# Gather update_ops from the first clone. These contain, for example,
# the updates for the batch_norm variables created by network_fn.
update_ops = tf.compat.v1.get_collection(
tf.compat.v1.GraphKeys.UPDATE_OPS, first_clone_scope)
# Add summaries for end_points.
end_points = clones[0].outputs
for end_point in end_points:
x = end_points[end_point]
summaries.add(tf.summary.histogram('activations/' + end_point, x))
summaries.add(
tf.summary.scalar('sparsity/' + end_point,
tf.nn.zero_fraction(x)))
# Add summaries for losses.
for loss in tf.compat.v1.get_collection(tf.compat.v1.GraphKeys.LOSSES,
first_clone_scope):
summaries.add(tf.summary.scalar('losses/%s' % loss.op.name, loss))
# Add summaries for variables.
for variable in slim.get_model_variables():
summaries.add(tf.summary.histogram(variable.op.name, variable))
#################################
# Configure the moving averages #
#################################
if moving_average_decay:
moving_average_variables = slim.get_model_variables()
variable_averages = tf.train.ExponentialMovingAverage(
moving_average_decay, global_step)
else:
moving_average_variables, variable_averages = None, None
#########################################
# Configure the optimization procedure. #
#########################################
with tf.device(deploy_config.optimizer_device()):
learning_rate = configure_learning_rate(dataset.num_samples,
global_step)
optimizer = configure_optimizer(learning_rate)
summaries.add(tf.summary.scalar('learning_rate', learning_rate))
if moving_average_decay:
# Update ops executed locally by trainer.
update_ops.append(
variable_averages.apply(moving_average_variables))
# Variables to train.
variables_to_train = get_variables_to_train(trainable_scopes)
# and returns a train_tensor and summary_op
total_loss, clones_gradients = model_deploy.optimize_clones(
clones, optimizer, var_list=variables_to_train)
# Add total_loss to summary.
summaries.add(tf.summary.scalar('total_loss', total_loss))
# Create gradient updates.
grad_updates = optimizer.apply_gradients(clones_gradients,
global_step=global_step)
update_ops.append(grad_updates)
update_op = tf.group(*update_ops)
with tf.control_dependencies([update_op]):
train_tensor = tf.identity(total_loss, name='train_op')
# Add the summaries from the first clone. These contain the summaries
# created by model_fn and either optimize_clones() or _gather_clone_loss().
summaries |= set(
tf.compat.v1.get_collection(tf.compat.v1.GraphKeys.SUMMARIES,
first_clone_scope))
# Merge all summaries together.
summary_op = tf.compat.v1.summary.merge_all()
config = tf.compat.v1.ConfigProto(allow_soft_placement=True,
log_device_placement=False)
config.gpu_options.allow_growth = True
config.gpu_options.visible_device_list = "0"
save_model_path = os.path.join(checkpoint_path, model_name,
"%s.ckpt" % model_name)
print(save_model_path)
# saver = tf.compat.v1.train.import_meta_graph('%s.meta'%save_model_path, clear_devices=True)
tf.compat.v1.disable_eager_execution()
# train the model
slim.learning.train(
train_op=train_tensor,
logdir=train_dir,
is_chief=(task == 0),
init_fn=_get_init_fn(save_model_path, train_dir=train_dir),
summary_op=summary_op,
number_of_steps=max_number_of_steps,
log_every_n_steps=log_every_n_steps,
save_summaries_secs=save_summaries_secs,
save_interval_secs=save_interval_secs,
# sync_optimizer=None,
session_config=config)
def main(_):
if not FLAGS.dataset_dir:
raise ValueError(
'You must supply the dataset directory with --dataset_dir')
tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.INFO)
with tf.Graph().as_default():
#######################
# Config model_deploy #
#######################
deploy_config = model_deploy.DeploymentConfig(
num_clones=FLAGS.num_clones,
clone_on_cpu=FLAGS.clone_on_cpu,
replica_id=FLAGS.task,
num_replicas=FLAGS.worker_replicas,
num_ps_tasks=FLAGS.num_ps_tasks)
# Create global_step
with tf.device(deploy_config.variables_device()):
global_step = slim.create_global_step()
######################
# Select the dataset #
######################
dataset = dataset_factory.get_dataset(FLAGS.dataset_name,
FLAGS.dataset_split_name,
FLAGS.dataset_dir)
######################
# Select the network #
######################
network_fn = nets_factory.get_network_fn(
FLAGS.model_name,
num_classes=(dataset.num_classes - FLAGS.labels_offset),
weight_decay=FLAGS.weight_decay,
is_training=True)
#####################################
# Select the preprocessing function #
#####################################
preprocessing_name = FLAGS.preprocessing_name or FLAGS.model_name
image_preprocessing_fn = preprocessing_factory.get_preprocessing(
preprocessing_name, is_training=True)
##############################################################
# Create a dataset provider that loads data from the dataset #
##############################################################
with tf.device(deploy_config.inputs_device()):
provider = slim.dataset_data_provider.DatasetDataProvider(
dataset,
num_readers=FLAGS.num_readers,
common_queue_capacity=20 * FLAGS.batch_size,
common_queue_min=10 * FLAGS.batch_size)
[image, label] = provider.get(['image', 'label'])
label -= FLAGS.labels_offset
train_image_size = FLAGS.train_image_size or network_fn.default_image_size
image = image_preprocessing_fn(image, train_image_size,
train_image_size)
images, labels = tf.compat.v1.train.batch(
[image, label],
batch_size=FLAGS.batch_size,
num_threads=FLAGS.num_preprocessing_threads,
capacity=5 * FLAGS.batch_size)
labels = slim.one_hot_encoding(
labels, dataset.num_classes - FLAGS.labels_offset)
batch_queue = slim.prefetch_queue.prefetch_queue(
[images, labels], capacity=2 * deploy_config.num_clones)
####################
# Define the model #
####################
def clone_fn(batch_queue):
"""Allows data parallelism by creating multiple clones of network_fn."""
images, labels = batch_queue.dequeue()
logits, end_points = network_fn(images)
#############################
# Specify the loss function #
#############################
if 'AuxLogits' in end_points:
slim.losses.softmax_cross_entropy(
end_points['AuxLogits'],
labels,
label_smoothing=FLAGS.label_smoothing,
weights=0.4,
scope='aux_loss')
slim.losses.softmax_cross_entropy(
logits,
labels,
label_smoothing=FLAGS.label_smoothing,
weights=1.0)
accuracy = slim.metrics.accuracy(
tf.cast(tf.argmax(input=logits, axis=1), dtype=tf.int32),
tf.cast(tf.argmax(input=labels, axis=1), dtype=tf.int32))
tf.compat.v1.add_to_collection('accuracy', accuracy)
end_points['train_accuracy'] = accuracy
return end_points
# Get accuracies for the batch
# Gather initial summaries.
summaries = set(
tf.compat.v1.get_collection(tf.compat.v1.GraphKeys.SUMMARIES))
clones = model_deploy.create_clones(deploy_config, clone_fn,
[batch_queue])
first_clone_scope = deploy_config.clone_scope(0)
# Gather update_ops from the first clone. These contain, for example,
# the updates for the batch_norm variables created by network_fn.
update_ops = tf.compat.v1.get_collection(
tf.compat.v1.GraphKeys.UPDATE_OPS, first_clone_scope)
# Add summaries for end_points.
end_points = clones[0].outputs
for end_point in end_points:
if 'accuracy' in end_point:
continue
x = end_points[end_point]
summaries.add(
tf.compat.v1.summary.histogram('activations/' + end_point, x))
summaries.add(
tf.compat.v1.summary.scalar('sparsity/' + end_point,
tf.nn.zero_fraction(x)))
train_acc = end_points['train_accuracy']
summaries.add(
tf.compat.v1.summary.scalar('train_accuracy',
end_points['train_accuracy']))
# Add summaries for losses.
for loss in tf.compat.v1.get_collection(tf.compat.v1.GraphKeys.LOSSES,
first_clone_scope):
summaries.add(
tf.compat.v1.summary.scalar('losses/%s' % loss.op.name, loss))
# @philkuz
# Add accuracy summaries
# TODO add if statemetn for n iterations
# images_val, labels_val= tf.train.batch(
# [image, label],
# batch_size=FLAGS.batch_size,
# num_threads=FLAGS.num_preprocessing_threads,
# capacity=5 * FLAGS.batch_size)
# # labels_val = slim.one_hot_encoding(
# # labels_val, dataset.num_classes - FLAGS.labels_offset)
# batch_queue_val = slim.prefetch_queue.prefetch_queue(
# [images_val, labels_val], capacity=2 * deploy_config.num_clones)
# logits, end_points = network_fn(images, reuse=True)
# # predictions = tf.nn.softmax(logits)
# predictions = tf.to_in32(tf.argmax(logits,1))
# logits_val, end_points_val = network_fn(images_val, reuse=True)
# predictions_val = tf.to_in32(tf.argmax(logits_val,1))
# labels_val = tf.squeeze(labels_val)
# labels = tf.squeeze(labels)
# names_to_values, names_to_updates = slim.metrics.aggregate_metric_map({
# 'train/accuracy': slim.metrics.streaming_accuracy(predictions, labels),
# 'val/accuracy': slim.metrics.streaming_accuracy(predictions_val, labels_val),
# })
# for metric_name, metric_value in names_to_values.items():
# op = tf.summary.scalar(metric_name, metric_value)
# # op = tf.Print(op, [metric_value], metric_name)
# summaries.add(op)
# Add summaries for variables.
# TODO something to remove some of these from tensorboard scalars
for variable in slim.get_model_variables():
summaries.add(
tf.compat.v1.summary.histogram(variable.op.name, variable))
#################################
# Configure the moving averages #
#################################
if FLAGS.moving_average_decay:
moving_average_variables = slim.get_model_variables()
variable_averages = tf.train.ExponentialMovingAverage(
FLAGS.moving_average_decay, global_step)
else:
moving_average_variables, variable_averages = None, None
#########################################
# Configure the optimization procedure. #
#########################################
with tf.device(deploy_config.optimizer_device()):
learning_rate = _configure_learning_rate(dataset.num_samples,
global_step)
optimizer = _configure_optimizer(learning_rate)
summaries.add(
tf.compat.v1.summary.scalar('learning_rate', learning_rate))
if FLAGS.sync_replicas:
# If sync_replicas is enabled, the averaging will be done in the chief
# queue runner.
optimizer = tf.compat.v1.train.SyncReplicasOptimizer(
opt=optimizer,
replicas_to_aggregate=FLAGS.replicas_to_aggregate,
total_num_replicas=FLAGS.worker_replicas,
variable_averages=variable_averages,
variables_to_average=moving_average_variables)
elif FLAGS.moving_average_decay:
# Update ops executed locally by trainer.
update_ops.append(
variable_averages.apply(moving_average_variables))
# Variables to train.
variables_to_train = _get_variables_to_train()
# and returns a train_tensor and summary_op
total_loss, clones_gradients = model_deploy.optimize_clones(
clones, optimizer, var_list=variables_to_train)
# Add total_loss to summary.
summaries.add(tf.compat.v1.summary.scalar('total_loss', total_loss))
# Create gradient updates.
grad_updates = optimizer.apply_gradients(clones_gradients,
global_step=global_step)
update_ops.append(grad_updates)
update_op = tf.group(*update_ops)
with tf.control_dependencies([update_op]):
train_tensor = tf.identity(total_loss, name='train_op')
# Add the summaries from the first clone. These contain the summaries
# created by model_fn and either optimize_clones() or _gather_clone_loss().
summaries |= set(
tf.compat.v1.get_collection(tf.compat.v1.GraphKeys.SUMMARIES,
first_clone_scope))
# Merge all summaries together.
summary_op = tf.compat.v1.summary.merge(list(summaries),
name='summary_op')
# @philkuz
# set the max_number_of_steps parameter if num_epochs is available
print('FLAGS.num_epochs', FLAGS.num_epochs)
if FLAGS.num_epochs is not None and FLAGS.max_number_of_steps is None:
FLAGS.max_number_of_steps = int(
FLAGS.num_epochs * dataset.num_samples / FLAGS.batch_size)
# FLAGS.max_number_of_steps = int(math.round(FLAGS.num_epochs / dataset.num_samples))
# setup the logdir
# @philkuz the train_dir setup
if FLAGS.experiment_name is not None:
experiment_dir = 'bs={},lr={},epochs={}/{}'.format(
FLAGS.batch_size, FLAGS.learning_rate, FLAGS.num_epochs,
FLAGS.experiment_name)
print(experiment_dir)
FLAGS.train_dir = os.path.join(FLAGS.train_dir, experiment_dir)
print(FLAGS.train_dir)
# @philkuz overriding train_step
def train_step(sess, train_op, global_step, train_step_kwargs):
"""Function that takes a gradient step and specifies whether to stop.
Args:
sess: The current session.
train_op: An `Operation` that evaluates the gradients and returns the
total loss.
global_step: A `Tensor` representing the global training step.
train_step_kwargs: A dictionary of keyword arguments.
Returns:
The total loss and a boolean indicating whether or not to stop training.
Raises:
ValueError: if 'should_trace' is in `train_step_kwargs` but `logdir` is not.
"""
start_time = time.time()
trace_run_options = None
run_metadata = None
should_acc = True # TODO make this not hardcoded @philkuz
if 'should_trace' in train_step_kwargs:
if 'logdir' not in train_step_kwargs:
raise ValueError(
'logdir must be present in train_step_kwargs when '
'should_trace is present')
if sess.run(train_step_kwargs['should_trace']):
trace_run_options = config_pb2.RunOptions(
trace_level=config_pb2.RunOptions.FULL_TRACE)
run_metadata = config_pb2.RunMetadata()
if not should_acc:
total_loss, np_global_step = sess.run(
[train_op, global_step],
options=trace_run_options,
run_metadata=run_metadata)
else:
total_loss, acc, np_global_step = sess.run(
[train_op, train_acc, global_step],
options=trace_run_options,
run_metadata=run_metadata)
time_elapsed = time.time() - start_time
if run_metadata is not None:
tl = timeline.Timeline(run_metadata.step_stats)
trace = tl.generate_chrome_trace_format()
trace_filename = os.path.join(
train_step_kwargs['logdir'],
'tf_trace-%d.json' % np_global_step)
tf.compat.v1.logging.info('Writing trace to %s',
trace_filename)
file_io.write_string_to_file(trace_filename, trace)
if 'summary_writer' in train_step_kwargs:
train_step_kwargs['summary_writer'].add_run_metadata(
run_metadata, 'run_metadata-%d' % np_global_step)
if 'should_log' in train_step_kwargs:
if sess.run(train_step_kwargs['should_log']):
if not should_acc:
tf.compat.v1.logging.info(
'global step %d: loss = %.4f (%.3f sec/step)',
np_global_step, total_loss, time_elapsed)
else:
tf.compat.v1.logging.info(
'global step %d: loss = %.4f train_acc = %.4f (%.3f sec/step)',
np_global_step, total_loss, acc, time_elapsed)
if 'should_stop' in train_step_kwargs:
should_stop = sess.run(train_step_kwargs['should_stop'])
else:
should_stop = False
return total_loss, should_stop
###########################
# Kicks off the training. #
###########################
slim.learning.train(
train_tensor,
logdir=FLAGS.train_dir,
train_step_fn=train_step,
master=FLAGS.master,
is_chief=(FLAGS.task == 0),
init_fn=_get_init_fn(),
summary_op=summary_op,
number_of_steps=FLAGS.max_number_of_steps,
log_every_n_steps=FLAGS.log_every_n_steps,
save_summaries_secs=FLAGS.save_summaries_secs,
save_interval_secs=FLAGS.save_interval_secs,
sync_optimizer=optimizer if FLAGS.sync_replicas else None)
