def q1():
global l_returnflag_group_size
global l_linestatus_group_size
returnflag_groups = np.unique(l_returnflag)
linestatus_groups = np.unique(l_linestatus)
l_returnflag_group_size = len(returnflag_groups)
l_linestatus_group_size = len(linestatus_groups)
inputs = [
tf.convert_to_tensor(l_shipdate, np.float32),
tf.convert_to_tensor(l_returnflag, np.float32),
tf.convert_to_tensor(l_linestatus, np.float32),
tf.convert_to_tensor(l_quantity, np.float32),
tf.convert_to_tensor(l_extendedprice, np.float32),
tf.convert_to_tensor(l_discount, np.float32),
tf.convert_to_tensor(l_tax, np.float32),
tf.convert_to_tensor(returnflag_groups, np.float32),
tf.convert_to_tensor(linestatus_groups, np.float32)
]
tpu_computation = tpu.rewrite(q1_computation, inputs)
tpu_grpc_url = TPUClusterResolver(
tpu=[os.environ['TPU_NAME']]).get_master()
with tf.Session(tpu_grpc_url) as sess:
sess.run(tpu.initialize_system())
sess.run(tf.global_variables_initializer())
for i in range(0, 5):
res = sess.run(tpu_computation)
sess.run(tpu.shutdown_system())
print(res)
return res
def __init__(self, iterations, hparams, per_host_v1=False):
tf.logging.info("TrainLowLevelRunner: constructor")
self.feature_structure = {}
self.loss = None
self.infeed_queue = []
self.enqueue_ops = []
self.dataset_initializer = []
self.is_local = ((hparams.master == "") and (hparams.tpu_name is None))
self.per_host_v1 = per_host_v1
self.iterations = iterations
self.sess = None
self.graph = tf.Graph()
self.hparams = hparams
with self.graph.as_default():
self.tpu_init = [tpu.initialize_system()]
self.tpu_shutdown = tpu.shutdown_system()
self.resolver = get_resolver(hparams)
session_config = tf.ConfigProto(allow_soft_placement=True,
isolate_session_state=True)
if self.hparams.tpu_name is None:
master = self.hparams.master
else:
cluster_spec = self.resolver.cluster_spec()
if cluster_spec:
session_config.cluster_def.CopyFrom(
cluster_spec.as_cluster_def())
master = self.resolver.get_master()
self.sess = tf.Session(master, graph=self.graph, config=session_config)
self.sess.run(self.tpu_init)
def __init__(self, iterations):
tf.logging.info("TrainLowLevelRunner: constructor")
self.feature_structure = {}
self.loss = None
self.infeed_queue = []
self.enqueue_ops = []
self.dataset_initializer = []
self.iterations = iterations
self.num_hosts = FLAGS.num_shards // FLAGS.num_shards_per_host
self.scaffold_fn = None
# Having two separate sessions and graphs to make the initialization faster.
self.input_sess = None
self.train_sess = None
self.input_graph = tf.Graph()
self.train_graph = None
self.tpu_cluster_resolver = tf.contrib.cluster_resolver.TPUClusterResolver(
FLAGS.tpu_name, zone=FLAGS.tpu_zone, project=FLAGS.gcp_project)
# Disable grappler for better performance.
self.session_config = tf.ConfigProto(
allow_soft_placement=True,
graph_options=tf.GraphOptions(
rewrite_options=rewriter_config_pb2.RewriterConfig(
disable_meta_optimizer=True)),
isolate_session_state=True)
cluster_spec = self.tpu_cluster_resolver.cluster_spec()
if cluster_spec:
self.session_config.cluster_def.CopyFrom(
cluster_spec.as_cluster_def())
self.tpu_init = [tpu.initialize_system()]
self.tpu_shutdown = tpu.shutdown_system()
self.init_sess = tf.Session(self.tpu_cluster_resolver.get_master(),
config=self.session_config)
self.init_sess.run(self.tpu_init)
self.queue = Queue.Queue()
def execute_tpu(self, graph_fn, inputs):
"""Constructs the graph, executes it on TPU and returns the result.
Args:
graph_fn: a callable that constructs the tensorflow graph to test. The
arguments of this function should correspond to `inputs`.
inputs: a list of numpy arrays to feed input to the computation graph.
Returns:
A list of numpy arrays or a scalar returned from executing the tensorflow
graph.
"""
with self.test_session(graph=tf.Graph()) as sess:
placeholders = [
tf.placeholder_with_default(v, v.shape) for v in inputs
]
tpu_computation = tpu.rewrite(graph_fn, placeholders)
sess.run(tpu.initialize_system())
sess.run([
tf.global_variables_initializer(),
tf.tables_initializer(),
tf.local_variables_initializer()
])
materialized_results = sess.run(tpu_computation,
feed_dict=dict(
zip(placeholders, inputs)))
sess.run(tpu.shutdown_system())
if (hasattr(materialized_results, '__len__')
and len(materialized_results) == 1
and (isinstance(materialized_results, list)
or isinstance(materialized_results, tuple))):
materialized_results = materialized_results[0]
return materialized_results
def __init__(self, iterations, train_steps):
tf.logging.info("TrainRunner: constructor")
self.feature_structure = {}
self.loss = None
self.infeed_queue = []
self.enqueue_ops = []
self.dataset_initializer = []
self.iterations = iterations
self.sess = None
self.input_sess = None
self.infeed_thread = None
if train_steps % iterations != 0:
train_steps = iterations * int(math.ceil(train_steps / iterations))
self.train_steps = train_steps
self.input_graph = tf.Graph()
tpu_init = [tpu.initialize_system()]
self.tpu_shutdown = tpu.shutdown_system()
self.cluster_resolver = tf.contrib.cluster_resolver.TPUClusterResolver(
FLAGS.tpu or FLAGS.master,
zone=FLAGS.tpu_zone,
project=FLAGS.gcp_project)
self.config = tf.ConfigProto(operation_timeout_in_ms=600 * 60 * 1000,
graph_options=tf.GraphOptions(
rewrite_options=rewriter_config_pb2.RewriterConfig(
disable_meta_optimizer=True)),
isolate_session_state=True)
cluster_spec = self.cluster_resolver.cluster_spec()
if cluster_spec:
self.config.cluster_def.CopyFrom(cluster_spec.as_cluster_def())
self.init_sess = tf.Session(self.cluster_resolver.get_master(), config=self.config)
self.init_sess.run(tpu_init)
def execute_tpu(self, graph_fn, inputs):
"""Constructs the graph, executes it on TPU and returns the result.
Args:
graph_fn: a callable that constructs the tensorflow graph to test. The
arguments of this function should correspond to `inputs`.
inputs: a list of numpy arrays to feed input to the computation graph.
Returns:
A list of numpy arrays or a scalar returned from executing the tensorflow
graph.
"""
with self.test_session(graph=tf.Graph()) as sess:
placeholders = [tf.placeholder_with_default(v, v.shape) for v in inputs]
tpu_computation = tpu.rewrite(graph_fn, placeholders)
sess.run(tpu.initialize_system())
sess.run([tf.global_variables_initializer(), tf.tables_initializer(),
tf.local_variables_initializer()])
materialized_results = sess.run(tpu_computation,
feed_dict=dict(zip(placeholders, inputs)))
sess.run(tpu.shutdown_system())
if (hasattr(materialized_results, '__len__') and
len(materialized_results) == 1 and
(isinstance(materialized_results, list) or
isinstance(materialized_results, tuple))):
materialized_results = materialized_results[0]
return materialized_results
def __init__(self, sess, use_tpu, mesh_shape, layout_rules):
super(MeshContext, self).__init__()
self._use_tpu = use_tpu
self._mesh_shape = mtf.convert_to_shape(mesh_shape)
self._layout_rules = layout_rules
self._d_assignment = None
self._num_hosts = None
self._num_cores = None
self._cpu_devices, self._gpu_devices = self._list_cpu_gpu_devices(sess)
if self._use_tpu:
topology = sess.run(tpu.initialize_system())
topo_object = tpu.Topology(serialized=topology)
self._num_cores = int(np.prod(topo_object.mesh_shape))
self._num_hosts = int(topo_object.num_tasks)
num_cores_per_host = int(self._num_cores // self._num_hosts)
assert num_cores_per_host == int(topo_object.num_tpus_per_task)
# Get a device_assignment object for mtf.
self._d_assignment = device_assignment.device_assignment(
topology, computation_shape=[1, 1, 1],
num_replicas=self._num_cores)
self._mesh_impl = mtf.simd_mesh_impl.SimdMeshImpl(
self._mesh_shape, self._layout_rules, None, self._d_assignment)
else:
self._mesh_impl = mtf.placement_mesh_impl.PlacementMeshImpl(
self._mesh_shape, self._layout_rules, self._gpu_devices)
def run(size):
a_ = []
b_ = []
c_ = []
for i in range(size):
a_.append((i * 1.0 + 4.0) * 2.5)
b_.append((i * 1.0 + 5.0) * 2.5)
c_.append((i * 1.0 + 6.0) * 0.1)
inputs = [tf.constant(a_), tf.constant(b_), tf.constant(c_)]
tpu_computation = tpu.rewrite(expression, inputs)
tpu_grpc_url = TPUClusterResolver(
tpu=[os.environ['TPU_NAME']]).get_master()
with tf.Session(tpu_grpc_url) as sess:
sess.run(tpu.initialize_system())
t1 = time()
sess.run(tf.global_variables_initializer())
sess.run(tpu_computation)
t2 = time()
print(str(size) + " : " + str(t2 - t1))
sess.run(tpu.shutdown_system())
print('Done !')
def run_graph(master, graph_spec, epoch):
"""Run graph_spec.graph with master."""
tf.logging.info("Running graph for epoch {}...".format(epoch))
with tf.Session(master, graph_spec.graph) as sess:
tf.logging.info("Initializing system for epoch {}...".format(epoch))
sess.run(
tpu.initialize_system(
embedding_config=graph_spec.embedding.config_proto))
tf.logging.info("Running before hook for epoch {}...".format(epoch))
graph_spec.hook_before(sess, epoch)
tf.logging.info("Running infeed for epoch {}...".format(epoch))
infeed_thread_fn = graph_spec.get_infeed_thread_fn(sess)
infeed_thread = threading.Thread(target=infeed_thread_fn)
tf.logging.info("Staring infeed thread...")
infeed_thread.start()
tf.logging.info("Running TPU loop for epoch {}...".format(epoch))
graph_spec.run_tpu_loop(sess, epoch)
tf.logging.info("Joining infeed thread...")
infeed_thread.join()
tf.logging.info("Running after hook for epoch {}...".format(epoch))
graph_spec.hook_after(sess, epoch)
def initialize_tpu(session=None, timeout_in_ms=None):
session = session or get_default_session()
with session.as_default():
op = tpu.initialize_system()
options = None
if timeout_in_ms:
options=config_pb2.RunOptions(timeout_in_ms=timeout_in_ms)
return session.run(op, options=options)
def main(unused_argv):
assert FLAGS.tpu_name
if FLAGS.tpu_name.startswith('grpc://'):
tpu_grpc_url = FLAGS.tpu_name
else:
tpu_cluster_resolver = contrib_cluster_resolver.TPUClusterResolver(
FLAGS.tpu_name, zone=None, project=None)
tpu_grpc_url = tpu_cluster_resolver.get_master()
sess = tf.Session(tpu_grpc_url)
with sess.graph.as_default():
contrib_tpu.initialize_system()
contrib_tpu.shutdown_system()
output_names = ['ConfigureDistributedTPU', 'ShutdownDistributedTPU']
model_def = tf.graph_util.convert_variables_to_constants(
sess, sess.graph.as_graph_def(), output_names)
print(model_def)
def __init__(self, iterations, train_steps=-1):
tf.logging.info("TrainRunner: constructor")
self.feature_structure = None
self.loss = None
self.infeed_queue = []
self.enqueue_ops = []
self.dataset_initializer = []
self.iterations = iterations
self.sess = None
self.input_sess = None
self.infeed_thread = None
if train_steps < 0:
train_steps = None
if train_steps is not None:
if train_steps % iterations != 0:
train_steps = iterations * int(
math.ceil(train_steps / iterations))
self.train_steps = train_steps
self.input_graph = tf.Graph()
with tf.Graph().as_default() as self.init_graph:
self.tpu_init = tpu.initialize_system()
self.tpu_shutdown = tpu.shutdown_system()
#self.cluster_resolver = tflex.TPUClusterResolver(
self.cluster_resolver = tf.contrib.cluster_resolver.TPUClusterResolver(
FLAGS.tpu or FLAGS.master,
zone=FLAGS.tpu_zone,
project=FLAGS.gcp_project)
self.config = tf.ConfigProto(
operation_timeout_in_ms=600 * 60 * 1000,
graph_options=tf.GraphOptions(
rewrite_options=rewriter_config_pb2.RewriterConfig(
disable_meta_optimizer=True)),
isolate_session_state=True)
cluster_spec = self.cluster_resolver.cluster_spec()
if cluster_spec:
self.config.cluster_def.CopyFrom(cluster_spec.as_cluster_def())
self.master = self.cluster_resolver.get_master()
self.init_sess = tf.Session(self.master,
graph=self.init_graph,
config=self.config)
tf.logging.info("TrainRunner: initializing TPU session...")
if not bool(int(os.environ.get('TPU_NO_INIT', '0'))):
tflex.run(self.init_sess, self.tpu_init)
tf.logging.info("TrainRunner: initializing TPU session (done)")
self.devices = self.init_sess.list_devices()
self.cores = sorted(
[x.name for x in self.devices if ':TPU:' in x.name])
self.num_cores = len(self.cores)
self.tpu_cores_per_host = 8
assert self.num_cores % self.tpu_cores_per_host == 0
self.num_hosts = self.num_cores // self.tpu_cores_per_host
print(self.config.cluster_def)
print('cores: %d hosts: %d ip: %s' %
(self.num_cores, self.num_hosts, self.master))
def __init__(self, iterations, train_steps, eval_steps):
tf.logging.info("TrainAndEvalRunner: constructor")
self.feature_structure = {}
self.eval_feature_structure = {}
self.loss = None
self.eval_loss = None
self.infeed_queue = []
self.eval_infeed_queue = []
self.enqueue_ops = []
self.num_hosts = FLAGS.num_cores // FLAGS.tpu_cores_per_host
self.dequeue_ops = []
self.queue = Queue.Queue()
self.eval_enqueue_ops = []
self.dataset_initializer = []
self.eval_dataset_initializer = []
self.iterations = iterations
self.steps_per_epoch = FLAGS.num_train_images // FLAGS.train_batch_size
self.iterator = None
self.sess = None
self.input_sess = None
self.eval_input_sess = None
self.eval_output_sess = None
self.infeed_thread = None
self.train_eval_thread = None
self.graph = tf.Graph()
self.input_graph = tf.Graph()
self.eval_input_graph = tf.Graph()
self.eval_output_graph = tf.Graph()
if train_steps % iterations != 0:
train_steps = iterations * int(math.ceil(train_steps / iterations))
self.train_steps = train_steps
self.max_train_iterations = self.train_steps // iterations
self.eval_steps = int(eval_steps)
self.eval_batch_size = FLAGS.eval_batch_size
tpu_init = [tpu.initialize_system()]
self.tpu_shutdown = tpu.shutdown_system()
self.tpu_cluster_resolver = tf.contrib.cluster_resolver.TPUClusterResolver(
FLAGS.tpu or FLAGS.master,
zone=FLAGS.tpu_zone,
project=FLAGS.gcp_project)
self.config = tf.ConfigProto(
operation_timeout_in_ms=600 * 60 * 1000,
allow_soft_placement=True,
graph_options=tf.GraphOptions(
rewrite_options=rewriter_config_pb2.RewriterConfig(
disable_meta_optimizer=True)),
isolate_session_state=True)
cluster_spec = self.tpu_cluster_resolver.cluster_spec()
if cluster_spec:
self.config.cluster_def.CopyFrom(cluster_spec.as_cluster_def())
self.master = self.tpu_cluster_resolver.get_master()
self.init_sess = tf.Session(self.master, config=self.config)
self.init_sess.run(tpu_init)
def filter_sum():
inputs = [tf.convert_to_tensor(l_quantity, np.float32)]
tpu_computation = tpu.rewrite(filter_sum_computation, inputs)
tpu_grpc_url = TPUClusterResolver(
tpu=[os.environ['TPU_NAME']]).get_master()
with tf.Session(tpu_grpc_url) as sess:
sess.run(tpu.initialize_system())
sess.run(tf.global_variables_initializer())
for i in range(0, 5):
res = sess.run(tpu_computation)
sess.run(tpu.shutdown_system())
print(res)
return res
def apply_comp(inputs):
tpu_computation = tpu.rewrite(apply, inputs)
tpu_grpc_url = TPUClusterResolver(
tpu=[os.environ['TPU_NAME']]).get_master()
with tf.Session(tpu_grpc_url) as sess:
sess.run(tpu.initialize_system())
sess.run(tf.global_variables_initializer())
t1 = time()
sess.run(tpu_computation)
t2 = time()
sess.run(tpu.shutdown_system())
print(t2 - t1)
def init_tpu(name, host=None, timeout_in_ms=600 * 60 * 1000):
tpu_init = [tpu.initialize_system()]
cluster_resolver = TPUClusterResolver(name, host=host)
config = tf.ConfigProto(operation_timeout_in_ms=timeout_in_ms,
graph_options=tf.GraphOptions(
rewrite_options=rewriter_config_pb2.RewriterConfig(
disable_meta_optimizer=True)),
isolate_session_state=True)
cluster_spec = cluster_resolver.cluster_spec()
if cluster_spec:
config.cluster_def.CopyFrom(cluster_spec.as_cluster_def())
init_sess = tf.Session(cluster_resolver.get_master(), config=config)
init_sess.run(tpu_init)
return init_sess, cluster_resolver
def run(tpu_computation, tpu_grpc_url):
reps = 1
times = []
for i in range(reps):
with tf.Session(tpu_grpc_url) as sess:
sess.run(tpu.initialize_system())
t1 = time()
sess.run(tf.global_variables_initializer())
sess.run(tpu_computation)
t2 = time()
print(str(i) + "_ : " + str(t2 - t1))
times.append(t2 - t1)
sess.run(tpu.shutdown_system())
print(sum(times) / reps)
def group_by():
unique_groups = np.unique(l_returnflag)
inputs = [
tf.convert_to_tensor(l_quantity, np.float32),
tf.convert_to_tensor(l_returnflag, np.float32),
tf.convert_to_tensor(unique_groups, np.float32)
]
tpu_computation = tpu.rewrite(group_by_computation, inputs)
tpu_grpc_url = TPUClusterResolver(
tpu=[os.environ['TPU_NAME']]).get_master()
with tf.Session(tpu_grpc_url) as sess:
sess.run(tpu.initialize_system())
sess.run(tf.global_variables_initializer())
for i in range(0, 5):
res = sess.run(tpu_computation)
sess.run(tpu.shutdown_system())
print(res)
def _run_tpu_computation(self):
"""Attempt to run computation graph directly on TPU."""
def _computation_fn(alpha, x, y):
return alpha * x + y
alpha = tf.Variable(3.0, name='alpha')
x = tf.Variable(tf.ones([3, 3], tf.float32), name='x')
y = tf.Variable(tf.ones([3, 3], tf.float32), name='y')
result = contrib_tpu.rewrite(_computation_fn, [alpha, x, y])
with tf.Session('grpc://{0}:8470'.format(self.tpu_ip)) as sess:
sess.run(contrib_tpu.initialize_system())
sess.run(tf.global_variables_initializer())
logging.info(sess.run(result))
sess.run(tpu.shutdown_system())
logging.info('Output should be a 3x3 matrix with all 4s.')
self.tpu_computation = 'Passed'
logging.info('Successfully ran a computation on the TPU')
def test_large_input(self):
if test_case.FLAGS.tpu_test:
input_size = 1408
min_level = 2
max_level = 6
batch_size = 2
num_boxes = 512
num_filters = 256
output_size = [7, 7]
with self.test_session() as sess:
features = []
for level in range(min_level, max_level + 1):
feat_size = int(input_size / 2**level)
features.append(
tf.constant(np.reshape(
np.arange(batch_size * feat_size * feat_size *
num_filters,
dtype=np.float32),
[batch_size, feat_size, feat_size, num_filters]),
dtype=tf.bfloat16))
boxes = np.array([
[[0, 0, 256, 256]] * num_boxes,
],
dtype=np.float32) / input_size
boxes = np.tile(boxes, [batch_size, 1, 1])
tf_boxes = tf.constant(boxes)
tf_levels = tf.random_uniform([batch_size, num_boxes],
maxval=5,
dtype=tf.int32)
def crop_and_resize_fn():
return spatial_ops.multilevel_roi_align(
features, tf_boxes, tf_levels, output_size)
tpu_crop_and_resize_fn = contrib_tpu.rewrite(
crop_and_resize_fn)
sess.run(contrib_tpu.initialize_system())
sess.run(tf.global_variables_initializer())
roi_features = sess.run(tpu_crop_and_resize_fn)
self.assertEqual(roi_features[0].shape,
(batch_size, num_boxes, output_size[0],
output_size[1], num_filters))
sess.run(contrib_tpu.shutdown_system())
def __init__(self, eval_steps):
tf.logging.info("EvalLowLevelRunner: constructor")
tf.logging.info("eval_steps: %s", eval_steps)
self.feature_structure = {}
self.infeed_queue = []
self.enqueue_ops = []
self.dataset_initializer = []
self.eval_steps = eval_steps
self.sess = None
self.eval_op = None
self.graph = tf.Graph()
self.outfeed_tensors = []
self.outfeed_names = []
self.dequeue_ops = {}
self.saver = None
self.tpu_cluster_resolver = None
with self.graph.as_default():
self.tpu_init = [tpu.initialize_system()]
self.tpu_shutdown = tpu.shutdown_system()
def timer(inputs):
reps = 2
times = []
for i in range(reps):
t1 = time()
tpu_computation = tpu.rewrite(blackscholes, inputs)
tpu_grpc_url = TPUClusterResolver(
tpu=[os.environ['TPU_NAME']]).get_master()
with tf.Session(tpu_grpc_url) as sess:
sess.run(tpu.initialize_system())
sess.run(tf.global_variables_initializer())
sess.run(tpu_computation)
sess.run(tpu.shutdown_system())
t2 = time()
print(str(i) + "_ : " + str(t2 - t1))
times.append(t2 - t1)
print(sum(times) / reps)
def __init__(self, input_fn, model_fn, params, num_steps):
self.feature_structure = {}
self.loss = None
self.enqueue_ops = None
self.metric_initializer = None
self.iterator = None
self.batch_size = params["batch_size"]
with tf.Graph().as_default() as self.graph:
self.build_model(params, input_fn, model_fn, num_steps)
self.tpu_init = tpu.initialize_system()
initializer = tf.global_variables_initializer()
self.tpu_shutdown = tpu.shutdown_system()
self.local_initializer = tf.local_variables_initializer()
self.saver = tf.train.Saver()
cluster_resolver = tf.contrib.cluster_resolver.TPUClusterResolver(
FLAGS.tpu or FLAGS.master)
self.sess = tf.Session(cluster_resolver.get_master(), graph=self.graph)
self.sess.run(self.tpu_init)
self.sess.run(initializer)
self.sess.run(self.local_initializer)
self.sess.run(self.iterator.initializer)
def execute_tpu_tf1(self, compute_fn, inputs, graph=None):
"""Executes compute_fn on TPU with Tensorflow 1.X.
Args:
compute_fn: a function containing Tensorflow computation that takes a list
of input numpy tensors, performs computation and returns output numpy
tensors.
inputs: a list of numpy arrays to feed input to the `compute_fn`.
graph: (optional) If not None, provided `graph` is used for computation
instead of a brand new tf.Graph().
Returns:
A list of numpy arrays or a single numpy array.
"""
with self.session(graph=(graph or tf.Graph())) as sess:
placeholders = [
tf.placeholder_with_default(v, v.shape) for v in inputs
]
def wrap_graph_fn(*args, **kwargs):
results = compute_fn(*args, **kwargs)
if (not (isinstance(results, dict)
or isinstance(results, tf.Tensor))
and hasattr(results, '__iter__')):
results = list(results)
return results
tpu_computation = contrib_tpu.rewrite(wrap_graph_fn, placeholders)
sess.run(contrib_tpu.initialize_system())
sess.run([
tf.global_variables_initializer(),
tf.tables_initializer(),
tf.local_variables_initializer()
])
materialized_results = sess.run(tpu_computation,
feed_dict=dict(
zip(placeholders, inputs)))
sess.run(contrib_tpu.shutdown_system())
return self.maybe_extract_single_output(materialized_results)
def __init__(self, eval_steps, hparams):
tf.logging.info("EvalLowLevelRunner: constructor")
tf.logging.info("eval_steps: %s", eval_steps)
self.feature_structure = {}
self.infeed_queue = []
self.enqueue_ops = []
self.dataset_initializer = []
self.is_local = ((hparams.master == "") and (hparams.tpu_name is None))
self.eval_steps = eval_steps
self.sess = None
self.eval_op = None
self.graph = tf.Graph()
self.hparams = hparams
self.outfeed_tensors = []
self.outfeed_names = []
self.dequeue_ops = {}
self.saver = None
with self.graph.as_default():
self.tpu_init = [tpu.initialize_system()]
self.tpu_shutdown = tpu.shutdown_system()
self.resolver = get_resolver(hparams)
session_config = tf.ConfigProto(allow_soft_placement=True,
operation_timeout_in_ms=600 * 60 *
1000) # 10 hours
if self.hparams.tpu_name is None:
master = self.hparams.master
else:
cluster_spec = self.resolver.cluster_spec()
if cluster_spec:
session_config.cluster_def.CopyFrom(
cluster_spec.as_cluster_def())
master = self.resolver.get_master()
self.sess = tf.Session(master, graph=self.graph, config=session_config)
self.sess.run(self.tpu_init)
def train_and_eval():
"""Trains and evaluates MeshTensorflow model without TPUEstimator.
TODO(lehou): Pack everything nicely as a set of APIs.
"""
tf.logging.info('FLAGS.master: {}'.format(FLAGS.master))
# Open a session to get the list of CPU devices to hold master variables.
with tf.Session(target=FLAGS.master,
config=tf.ConfigProto(allow_soft_placement=True)) as sess:
topology = sess.run(tpu.initialize_system())
cpu_devices = _list_cpu_devices(sess)
topo_object = tf.contrib.tpu.Topology(serialized=topology)
num_cores = int(np.prod(topo_object.mesh_shape))
num_hosts = int(topo_object.num_tasks)
num_cores_per_host = int(num_cores // num_hosts)
assert num_cores_per_host == int(topo_object.num_tpus_per_task)
# Get a device_assignment object for mtf.
d_assignment = device_assignment.device_assignment(
topology, computation_shape=[1, 1, 1], num_replicas=num_cores)
# Get mesh_impl.
mesh_shape = mtf.convert_to_shape(FLAGS.mesh_shape)
layout_rules = unet.get_layout()
mesh_impl = mtf.simd_mesh_impl.SimdMeshImpl(mesh_shape, layout_rules, None,
d_assignment)
for _ in range(FLAGS.num_training_loops):
_train_phase(mesh_impl, cpu_devices, d_assignment, num_hosts,
num_cores)
_eval_phase(mesh_impl, cpu_devices, d_assignment, num_hosts, num_cores)
_shutdown()
tf.logging.info('finished.')
def run_monitored_session(cross_entropy, log_dir, required_steps, class_range,
save_checkpoint_steps, validation_steps,
train_step,
augmentation_info, device,
training_nn_params, training_tensor,
testing_nn_params, testing_tensor,
validation_nn_params, validation_tensor):
read_op_value = None
augmentation_restorer = None
if augmentation_info.perform_shadow_augmentation:
if augmentation_info.shadow_struct is not None or augmentation_info.shadow_struct.shadow_op_initializer is not None:
augmentation_restorer = augmentation_info.shadow_struct.shadow_op_creater()
# Ready ops are overriden, as default ready ops awaits all variables to be initialized
# but actually some of the variables(such as cycle-gan graphs) are not initialized but restored
read_op_value = constant([])
is_gpu_or_cpu = (device == "gpu" or device == "cpu")
if is_gpu_or_cpu:
config = tf.ConfigProto(allow_soft_placement=True, log_device_placement=False)
set_all_gpu_config()
master = ''
else:
config = None
tpu_worker = 'grpc://' + os.environ['COLAB_TPU_ADDR']
# master = TPUClusterResolver(tpu=tpu_worker).get_master()
master = tpu_worker
print("TPU master")
print(master)
validation_hook = ValidationHook(validation_nn_params, validation_tensor, class_range, required_steps,
validation_steps,
log_dir)
test_iteration_count = 100
test_hook = TestHook(testing_nn_params, testing_tensor, cross_entropy, test_iteration_count, class_range)
initializer_hook = InitializerHook(training_nn_params, training_tensor, augmentation_info, augmentation_restorer)
stop_on_step_hook = StopAtStepHook(last_step=required_steps - 1)
nan_tensor_hook = NanTensorHook(loss_tensor=cross_entropy, fail_on_nan_loss=False)
hooks = [initializer_hook,
validation_hook,
test_hook,
stop_on_step_hook,
nan_tensor_hook]
if is_gpu_or_cpu:
# Only restore nn core variables along with the optimizer and global step variables
nn_core_restorer = tf.train.Saver(
max_to_keep=20,
var_list=slim.get_variables_to_restore(include=["nn_core"]) +
slim.get_variables_to_restore(include=["global_step"]) +
slim.get_variables_to_restore(include=["training_optimizer"]), name="nn_core_restorer")
training_scaffold = Scaffold(saver=nn_core_restorer,
ready_for_local_init_op=read_op_value,
ready_op=read_op_value)
session = tf.train.MonitoredTrainingSession(master=master,
checkpoint_dir=log_dir,
summary_dir=log_dir,
config=config, is_chief=True,
save_summaries_steps=test_iteration_count,
save_checkpoint_steps=save_checkpoint_steps,
scaffold=training_scaffold,
hooks=hooks)
# session = LocalCLIDebugWrapperSession(session)
with session as monitored_sess:
while not monitored_sess.should_stop():
monitored_sess.run([train_step])
else:
session = tf.Session(target=master, config=config)
session.run(tpu.initialize_system())
session.run(tf.group(tf.global_variables_initializer(),
tf.local_variables_initializer()))
initializer_hook.after_create_session(session, None)
while session.run(test_hook._global_step_tensor) < required_steps:
try:
session.run(train_step)
test_hook.after_run_with_session(session)
except tf.errors.OutOfRangeError:
break
validation_hook.end(session)
session.run(tpu.shutdown_system())
session.close()
result = TrainingResult(validation_accuracy=validation_hook.validation_accuracy,
test_accuracy=test_hook.testing_accuracy, loss=test_hook.loss)
return result
def __init__(self,
iterations,
num_cores_per_shard=1,
input_partition_dims=None):
tf.logging.info("TrainLowLevelRunner: constructor")
self.feature_structure = {}
self.loss = None
self.infeed_queue = []
self.enqueue_ops = []
self.dataset_initializer = []
self.iterations = iterations
# TODO(wangtao): change FLAGS.num_shards_per_host to
# FLAGS.num_cores_per_host after other low level API
# support spatial partition. FLAGS.num_shards_per_host means number of TPU
# cores for each host.
self.replicas_per_worker = FLAGS.num_shards_per_host // num_cores_per_shard
self.num_hosts = FLAGS.num_shards * num_cores_per_shard // FLAGS.num_shards_per_host
self.num_shards = FLAGS.num_shards
self.scaffold_fn = None
# Having two separate sessions and graphs to make the initialization faster.
self.input_sess = None
self.train_sess = None
self.input_graph = tf.Graph()
self.train_graph = None
self.tpu_cluster_resolver = tf.contrib.cluster_resolver.TPUClusterResolver(
FLAGS.tpu_name, zone=FLAGS.tpu_zone, project=FLAGS.gcp_project)
# Disable grappler for better performance.
self.session_config = tf.ConfigProto(
allow_soft_placement=True,
graph_options=tf.GraphOptions(
rewrite_options=rewriter_config_pb2.RewriterConfig(
disable_meta_optimizer=True)),
isolate_session_state=True)
cluster_spec = self.tpu_cluster_resolver.cluster_spec()
if cluster_spec:
self.session_config.cluster_def.CopyFrom(
cluster_spec.as_cluster_def())
self.tpu_init = tpu.initialize_system()
self.tpu_shutdown = tpu.shutdown_system()
self.init_sess = tf.Session(self.tpu_cluster_resolver.get_master(),
config=self.session_config)
self.queue = Queue.Queue()
# Init for spatial partitioning.
self.device_topology = self.init_sess.run(self.tpu_init)
self.input_partition_dims = input_partition_dims
self.use_spatial_partition = (
input_partition_dims is not None
and int(np.prod(FLAGS.input_partition_dims)) > 1)
self.num_cores_per_shard = num_cores_per_shard
if self.use_spatial_partition:
computation_shape = _NUM_CORES_TO_COMPUTATION_SHAPE[
self.num_cores_per_shard]
self.device_assignment = tpu_device_assignment.device_assignment(
topology=self.device_topology,
computation_shape=computation_shape,
num_replicas=self.num_shards)
tf.logging.info("num_cores_per_shard: %d",
self.num_cores_per_shard)
tf.logging.info("num_hosts: %d", self.num_hosts)
tf.logging.info("replicas_per_worker: %d",
self.replicas_per_worker)
tf.logging.info("computation_shape: %s", str(computation_shape))
tf.logging.info("num_shards: %d", self.num_shards)
tf.logging.info(
"device_assignment.topology.device_coordinates: %s",
str(self.device_assignment.topology.device_coordinates))
tf.logging.info("device_assignment.core_assignment: %s",
str(self.device_assignment.core_assignment))
else:
self.device_assignment = None
def tf_train_flow(
train_once_fn,
model_dir=None,
log_dir=None,
max_models_keep=1,
save_interval_seconds=600,
save_interval_steps=1000,
num_epochs=None,
num_steps=None,
save_model=True,
save_interval_epochs=None,
freeze_graph=False,
num_steps_per_epoch=0,
restore_from_latest=True,
metric_eval_fn=None,
valid_interval_epochs=0,
inference_fn=None,
inference_interval_epochs=0,
init_fn=None,
restore_fn=None,
restore_include=None,
restore_exclude=None,
save_all_scope=False, #TODO save load from restore scope only but svae all
variables_to_restore=None,
variables_to_save=None, #by default will be the same as variables_to_restore
output_collection_names=None,
output_node_names=None,
learning_rate=None, #not use yet, just use in train_once
learning_rate_patience=None,
learning_rate_decay_factor=None,
write_during_train=True,
model=None,
sess=None):
"""
similary flow as tf_flow, but add model try reload and save
"""
use_horovod = 'OMPI_COMM_WORLD_RANK' in os.environ
model_dir_ = model_dir
if use_horovod and hvd.rank() != 0:
model_dir = None
if sess is None:
#TODO melt.get_session is global session but may cause non close at last
sess = melt.get_session()
if FLAGS.use_tpu:
sess.run(tpu.initialize_system())
#logging.info('tf_train_flow start')
#logging.info('max_models_keep:', max_models_keep)
#logging.info('save_interval_seconds:', save_interval_seconds)
if model_dir:
if model:
checkpoint = tf.train.Checkpoint(model=model)
ckpt_dir = model_dir + '/ckpt'
checkpoint_prefix = os.path.join(ckpt_dir, 'ckpt')
#this is usefull for you use another model with another scope, and just load and restore/save initalize your scope vars!
#this is not for finetune but mainly for like using another model as in predict like this introducing graph other model scope and ignore here
# var_list = None if not restore_scope else tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope=restore_scope)
# #logging.info('-------------var_list', var_list)
# if not variables_to_restore:
# variables_to_restore = var_list
if not variables_to_restore:
variables_to_restore = slim.get_variables_to_restore(
include=restore_include, exclude=restore_exclude)
if not variables_to_save:
variables_to_save = variables_to_restore
if save_all_scope:
variables_to_save = None
#if variables_to_restore is None:
logging.info('variables_to_restore from %s' % model_dir)
#load all var in checkpoint try to save all var(might more then original checkpoint) if not specifiy variables_to_save
varnames_in_checkpoint = melt.get_checkpoint_varnames(model_dir)
#logging.info('varnames_in_checkpoint: {}'.format(varnames_in_checkpoint))
# TODO has someproblem say tf.Variable 'r_net/text_encoder/cudnn_rnn/cu_dnngru/recurrent_kernel/adam_v:0' even though in checkpoint I have renated it as ignore/rnet
variables_to_restore_from_model = slim.get_variables_to_restore(
include=varnames_in_checkpoint)
#logging.info('variables_to_restore_from_model: {}'.format(variables_to_restore_from_model))
if not variables_to_restore:
variables_to_restore = variables_to_restore_from_model
else:
variables_to_restore = [
v for v in variables_to_restore
if v in variables_to_restore_from_model
]
if restore_exclude:
for excl in restore_exclude:
variables_to_restore = [
v for v in variables_to_restore if not excl in v.name
]
#--tf 1.6 adadelta will have same vars...
variables_to_restore = list(set(variables_to_restore))
#logging.info('variables_to_restore', variables_to_restore[:100])
logging.info('variables_to_restore', [
x for x in variables_to_restore if not 'OptimizeLoss' in x.name
][:100])
##finally remove global_step since melt.apps.train will handle it!
global_step = tf.train.get_or_create_global_step()
#variables_to_restore = [v for v in variables_to_restore if not tf.GraphKeys.GLOBAL_STEP in v.name]
#variables_to_restore = [v for v in variables_to_restore if not 'learning_rate' in v.name]
# TODO fixme if step, step2.. and in checkpoint step then here will be step2...
#print('------------', [v for v in variables_to_restore if 'step' in v.name])
loader = tf.train.Saver(var_list=variables_to_restore)
logging.info('max models to keep {}, keep every {} hours'.format(
max_models_keep, save_interval_seconds / 3600.0))
saver = tf.train.Saver(
max_to_keep=max_models_keep,
keep_checkpoint_every_n_hours=save_interval_seconds / 3600.0,
var_list=variables_to_save)
epoch_saver = tf.train.Saver(var_list=variables_to_save, max_to_keep=1000)
best_epoch_saver = tf.train.Saver(var_list=variables_to_save)
#logging.info('variables_to_save:{}'.format(variables_to_save))
# # #TODO for safe restore all init will be ok ?
# # if variables_to_restore is None:
init_op = tf.group(
tf.global_variables_initializer(
), #variables_initializer(global_variables())
tf.local_variables_initializer()
) #variables_initializer(local_variables())
# # else:
# # init_op = tf.group(tf.variables_initializer(variables_to_restore),
# # tf.local_variables_initializer())
##--mostly this will be fine except for using assistant predictor, initialize again! will make assistant predictor wrong
##so assume to all run init op! if using assistant predictor, make sure it use another session
# https://stackoverflow.com/questions/35164529/in-tensorflow-is-there-any-way-to-just-initialize-uninitialised-variables
# def guarantee_initialized_variables(session, list_of_variables = None):
# if list_of_variables is None:
# list_of_variables = tf.global_variables()
# uninitialized_variables = list(tf.get_variable(name) for name in
# session.run(tf.report_uninitialized_variables(list_of_variables)))
# return unintialized_variables
# unintialized_variables = guarantee_initialized_variables(sess)
# init_op = tf.group(tf.initialize_variables(uninitialized_vars), tf.local_variables_initializer())
timer = gezi.Timer('sess run init_op in melt.tf_train_flow')
#model.save('./weights')
# notice
sess.run(init_op)
timer.print_elapsed()
#melt.init_uninitialized_variables(sess)
#pre_step means the step last saved, train without pretrained,then -1
pre_step = -1
fixed_pre_step = -1 #fixed pre step is for epoch num to be correct if you change batch size
#print(model_dir)
pre_epoch = None
if model_dir:
model_path = _get_model_path(model_dir, save_model)
# if not model_path:
# model_path = _get_model_path(os.path.join(model_dir, 'epoch'))
#print(model_path)
model_dir = gezi.get_dir(
model_dir) #incase you pass ./model/model-ckpt1000 -> ./model
if model_path is not None:
if not restore_from_latest:
logging.info('using recent but not latest model')
model_path = melt.recent_checkpoint(model_dir)
model_name = os.path.basename(model_path)
timer = gezi.Timer(
'Loading and training from existing model [%s]' % model_path)
if restore_fn is not None:
restore_fn(sess)
loader.restore(sess, model_path)
## not supported
#model.save()
#model.save_weights('./weights')
timer.print()
#pre_step = melt.get_model_step(model_path) - 1 if FLAGS.global_step is None else FLAGS.global_step -1
# TODO check ..
pre_step = sess.run(tf.train.get_global_step()) - 1
pre_epoch = melt.get_model_epoch(
model_path
) if FLAGS.global_epoch is None else FLAGS.global_epoch
fixed_pre_step = pre_step
# if pre_epoch is not None:
# #like using batch size 32, then reload train using batch size 64
# if abs(pre_step / num_steps_per_epoch - pre_epoch) > 0.1:
# fixed_pre_step = int(pre_epoch * num_steps_per_epoch)
# logging.info('Warning, epoch is diff with pre_step / num_steps_per_epoch:{}, pre_epoch:{},maybe you change batch size and we will adjust to set pre_step as {}'\
# .format(pre_step / num_steps_per_epoch, pre_epoch, fixed_pre_step))
else:
latest_checkpoint = None
if not use_horovod: #now will hang
try:
latest_checkpoint = tf.train.latest_checkpoint(ckpt_dir)
if latest_checkpoint:
logging.info(
'Try start from eager trained mode, latest checkpoint:',
latest_checkpoint)
checkpoint.restore(latest_checkpoint).run_restore_ops(
session=sess)
pre_epoch = int(latest_checkpoint.split('-')[-1])
#pre_step = pre_epoch * num_steps_per_epoch - 1
# TODO check
pre_step = sess.run(tf.train.get_global_step()) - 1
fixed_pre_step = pre_step
logging.info('Start step is:', pre_step)
except Exception:
logging.info(
'Something wrong with restore from eager trained model'
)
if latest_checkpoint is None:
logging.info('Train all start step 0')
#https://stackoverflow.com/questions/40220201/tensorflow-tf-initialize-all-variables-vs-tf-initialize-local-variables
#tf.initialize_all_variables() is a shortcut to tf.initialize_variables(tf.all_variables()),
#tf.initialize_local_variables() is a shortcut to tf.initialize_variables(tf.local_variables()),
#which initializes variables in GraphKeys.VARIABLES and GraphKeys.LOCAL_VARIABLE collections, respectively.
#init_op = tf.group(tf.global_variables_initializer(),
# tf.local_variables_initializer())
#[var for var in tf.all_variables() if var.op.name.startswith(restore_scope)] will be the same as tf.get_collection(tf.GraphKeys.GLOBAL_VARIABLES, scope=restore_scope)
#sess.run(init_op)
#like use image model, build image graph, reload first train, and then will go to same checkpoint all varaible just restore will ok
#for finetune from loading other model init
if init_fn is not None:
init_fn(sess)
if gezi.env_has('METRIC'):
l = metric_eval_fn(model_path)
print(list(zip(l[1], l[0])))
exit(0)
#sess.run(tf.assign(global_step, tf.constant(global_step_val, dtype=tf.int64)))
try:
learning_rate = tf.get_collection('learning_rate')[-1]
learning_rate_weight = tf.get_collection('learning_rate_weight')[-1]
sess.run(tf.assign(learning_rate,
learning_rate * learning_rate_weight))
except Exception:
# if not using weight_decay but using optimizer decay then will go here as learning rate is a tensor can not assign
pass
try:
logging.info('Actual start global step:',
sess.run(global_step), 'learning rate:',
sess.run(learning_rate), 'learning_rate_weight:',
sess.run(learning_rate_weight))
except Exception:
pass
if model_dir_:
#if save_interval_epochs and num_steps_per_epoch and num_steps >= 0:
epoch_dir = os.path.join(model_dir_, 'epoch')
gezi.try_mkdir(epoch_dir)
checkpoint_path = os.path.join(model_dir_, 'model.ckpt')
coord = tf.train.Coordinator()
threads = tf.train.start_queue_runners(sess=sess, coord=coord)
if use_horovod:
bcast = hvd.broadcast_global_variables(0)
sess.run(bcast)
#tf.train.write_graph(sess.graph_def, model_dir, 'train.pbtxt')
only_one_step = False
try:
if use_horovod:
## TODO FIXME why bcast here not work ? simple test work see tests/bcast.py
#comm.bcast(pre_step, root=0)
temp = np.array([pre_step, fixed_pre_step])
comm.Bcast(temp, root=0)
pre_step = temp[0]
fixed_pre_step = temp[1]
step = start = pre_step + 1
fixed_step = fixed_pre_step + 1
#first = True
#hack just for save one model after load
if num_steps < 0 or (num_steps and num_steps < step):
logging.info('just load and resave then exit')
model_path_ = _get_checkpoint_path(checkpoint_path,
step,
num_steps_per_epoch,
epoch=pre_epoch)
saver.save(sess, model_path_, global_step=step + 1)
if freeze_graph:
melt.freeze_graph(sess, model_path_, step + 1,
output_collection_names, output_node_names)
sess.close()
exit(0)
if num_epochs < 0:
only_one_step = True
logging.info('just run one step')
if FLAGS.work_mode != 'train':
assert not os.path.isdir(FLAGS.model_dir), FLAGS.model_dir
if 'valid' in FLAGS.work_mode:
vals, names = metric_eval_fn(FLAGS.model_dir)
logging.info(list(zip(names, vals)))
if 'test' in FLAGS.work_mode:
inference_fn(FLAGS.model_dir)
exit(0)
#early_stop = True #TODO allow config
num_bad_epochs = 0
pre_epoch_eval_loss = 1e20
best_epoch_eval_loss = 1e20
num_allowed_bad_epochs = 4 #allow 5 non decrease eval loss epochs before stop
epoch_saved_step = 0
while not coord.should_stop():
model_step_path = None
if model_dir_:
model_path_ = os.path.join(
epoch_dir, 'model.ckpt-%.2f' %
(fixed_step / float(num_steps_per_epoch)))
model_step_path_ = model_path_ + '-' + str(step)
if (write_during_train and metric_eval_fn is not None
and valid_interval_epochs and fixed_step %
int(num_steps_per_epoch * valid_interval_epochs) == 0):
model_step_path = model_step_path_
else:
model_step_path = None
if step == 0:
model_step_path = None
#print('--------------------step', step)
stop = train_once_fn(
sess,
step,
is_start=(step == start),
fixed_step=fixed_step,
num_epochs=num_epochs,
model_path=model_step_path,
use_horovod=use_horovod,
## TODO FIXME this line will cause tensorflow.python.framework.errors_impl.NotFoundError: Resource localhost/save_counter/N10tensorflow3VarE does not exist.
)
#first = False
if only_one_step:
stop = True
step += 1
fixed_step += 1
if save_model and step and model_dir:
#step 0 is also saved! actually train one step and save
if step % save_interval_steps == 0:
timer = gezi.Timer(
'save model step %d to %s' % (step, checkpoint_path),
False)
model_path_ = _get_checkpoint_path(checkpoint_path,
fixed_step,
num_steps_per_epoch)
saver.save(sess, model_path_, global_step=step)
if freeze_graph:
melt.freeze_graph(sess, model_path_, step,
output_collection_names,
output_node_names)
#if log_dir != model_dir:
# assert log_dir
# command = 'rsync -l -r -t %s/* %s' % (log_dir, model_dir)
# print(command, file=sys.stderr)
# os.system(command)
timer.print_elapsed()
if save_interval_steps and num_steps_per_epoch and fixed_step % int(
num_steps_per_epoch * save_interval_epochs) == 0:
# TODO only epoch in name not sep ?
epoch_saved_step = step
model_path_ = os.path.join(
epoch_dir, 'model.ckpt-%.2f' %
(fixed_step / float(num_steps_per_epoch)))
model_step_path = model_path_ + '-' + str(step)
epoch_saver.save(sess, model_path_, global_step=step)
#epoch_saver.save(sess, model_path_)
## TODO FIXME do not support tf.keras save currently with horovod
# if model:
# #model.save_weights(epoch_dir + '/ckpt-%.2f' % (fixed_step / float(num_steps_per_epoch)))
# # TODO FIXME if restart will save from 1... again..
# checkpoint.save(checkpoint_prefix, session=sess)
# #print(sess.run(checkpoint.save_counter))
if freeze_graph:
melt.freeze_graph(sess, model_path_, step,
output_collection_names,
output_node_names)
if write_during_train:
if inference_fn is not None and inference_interval_epochs and fixed_step % int(
num_steps_per_epoch *
inference_interval_epochs) == 0:
model_step_path = model_path_ + '-' + str(step)
try:
#print('--------------inference fn')
inference_fn(model_path=model_step_path)
except Exception:
logging.info(traceback.format_exc())
# if metric_eval_fn is not None and valid_interval_epochs and fixed_step % int(num_steps_per_epoch * valid_interval_epochs) == 0:
# model_step_path = model_path_ + '-' + str(step)
# try:
# metric_eval_fn(model_path=model_step_path)
# except Exception:
# logging.info(traceback.format_exc())
if stop is True:
print('Early stop running %d stpes' % (step), file=sys.stderr)
raise tf.errors.OutOfRangeError(
None, None, 'Early stop running %d stpes' % (step))
if num_steps and (step + 1) == start + num_steps:
raise tf.errors.OutOfRangeError(None, None,
'Reached max num steps')
#max_num_epochs = 1000
max_num_epochs = num_epochs
#if max_num_epochs and num_steps_per_epoch and fixed_step // num_steps_per_epoch >= max_num_epochs:
if max_num_epochs and num_steps_per_epoch and fixed_step / num_steps_per_epoch > max_num_epochs:
raise tf.errors.OutOfRangeError(
None, None,
'Reached max num epochs of %d' % max_num_epochs)
#except tf.errors.OutOfRangeError, e:
except tf.errors.OutOfRangeError:
# if run 2 epoch and we have just epoch saved, do not need to save only 1 step more model
if (step - epoch_saved_step > 1) and not (
step == start
) and save_model and step % save_interval_steps != 0 and model_dir:
model_path_ = _get_checkpoint_path(checkpoint_path, step,
num_steps_per_epoch)
saver.save(sess, model_path_, global_step=step)
if freeze_graph:
melt.freeze_graph(sess, model_path_, step,
output_collection_names, output_node_names)
if log_dir != model_dir:
assert log_dir
command = 'rsync -l -r -t %s/* %s' % (log_dir, model_dir)
print(command, file=sys.stderr)
os.system(command)
if only_one_step:
logging.info('Done one step')
exit(0)
# if (step - epoch_saved_step > 1) and metric_eval_fn is not None:
# metric_eval_fn(model_path=model_step_path)
if (num_epochs and fixed_step / num_steps_per_epoch >= num_epochs) or (
num_steps and step == start + num_steps):
logging.info('Done training for %.3f epochs, %d steps.' %
(fixed_step / num_steps_per_epoch, step))
#FIXME becase coord.join seems not work, RuntimeError: Coordinator stopped with threads still running: Thread-9
exit(0)
else:
logging.info('Should not stop, but stopped at epoch: %.3f' %
(fixed_step / num_steps_per_epoch))
logging.info(traceback.format_exc())
#raise e
finally:
coord.request_stop()
coord.join(threads, stop_grace_period_secs=5)
#FIMXE due to use melt.get_session(global not handle del well)
#Done training for 3090020 steps.
#Exception TypeError: "'NoneType' object is not callable" in <bound method Session.__del__ of <tensorflow.python.client.session.Session object at 0x7f6cf33cd450>> ignored
if FLAGS.use_tpu:
sess.run(tpu.shutdown_system())
sess.close()
def flops():
x = tf.random_uniform([N, N])
y = tf.random_uniform([N, N])
def _matmul(x, y):
return tf.tensordot(x, y, axes=[[1], [0]]), y
return tf.reduce_sum(tpu.repeat(COUNT, _matmul, [x, y]))
tpu_ops = tpu.batch_parallel(flops, [], num_shards=8)
session = tf.Session(tpu_cluster)
try:
print('Warming up...')
session.run(tpu.initialize_system())
session.run(tpu_ops)
print('Profiling')
start = time.time()
session.run(tpu_ops)
end = time.time()
elapsed = end - start
print(elapsed,
'TFlops: {:.2f}'.format(1e-12 * 8 * COUNT * 2 * N * N * N / elapsed))
except Exception as e:
print(e)
finally:
session.run(tpu.shutdown_system())
session.close()
def get_initialization_ops(self): return [tpu.initialize_system()]
