def ConstructFPropBPropGraph(self):
# We need to override this since constructing the BPropGraph
# creates slot variables.
with py_utils.OpportunisticVariableReuseScope():
with py_utils.VariableRenameScope(
self.params.variable_renaming_rules):
super().ConstructFPropBPropGraph()
def _MaybeConstructSharedModel(self, train_cfg):
"""Construct a single shared copy of the model if this is a MultiTaskModel.
For MultiTaskModels, we create a MultiTaskSubModel for each task, but
construct the model only once.
Args:
train_cfg: The params for a SingleTaskModel or MultiTaskModel.
Returns:
A MultiTaskModel, if train_cfg is a MultiTaskModel params object.
"""
if not issubclass(train_cfg.cls, base_model.MultiTaskModel):
return None
with self._graph.as_default(), tf.container(self._container_id):
with self._cluster, tf.device(
self._cluster.job_spec.name if not FLAGS.
cluster_placer_in_executor else self._cluster.GetPlacer()):
with py_utils.VariableRenameScope(
self._variable_renaming_rules):
_ = py_utils.GetOrCreateGlobalStepVar()
shared_model = train_cfg.Instantiate()
shared_model.InstantiateVariables()
return shared_model
def _MaybeConstructSharedModel(self, train_cfg):
"""Construct a single shared copy of the model if this is a MultiTaskModel.
If the share_model_object parameter is set, for MultiTaskModels,
we create a MultiTaskSubModel for each task, but construct the model only
once.
Args:
train_cfg: The params for a SingleTaskModel or MultiTaskModel.
Returns:
A MultiTaskModel, if train_cfg is a MultiTaskModel params object.
"""
if not issubclass(train_cfg.cls, base_model.MultiTaskModel):
return None
if not train_cfg.share_model_object:
return None
with self._cluster, tf.container(
self._container_id), contextlib.ExitStack() as stack:
if not py_utils.IsEagerMode():
stack.enter_context(self._graph.as_default())
stack.enter_context(tf.device(self._cluster.GetPlacer()))
with py_utils.VariableStore(), py_utils.VariableRenameScope(
self._variable_renaming_rules):
py_utils.GetOrCreateGlobalStepVar()
shared_model = train_cfg.Instantiate()
return shared_model
def __init__(self, params):
super().__init__(params)
p = params
if p.input_symbols:
assert p.input_symbols.num_symbols() == p.input_vocab_size
if p.input_symbols:
assert p.output_symbols.num_symbols() == p.output_vocab_size
if p.share_embeddings:
renames = [("(.*)/token_emb/(.*)", "%s/shared_emb/token_emb/%s")]
else:
renames = [("(.*)/(?:encoder|spell_encoder)/token_emb/(.*)",
"%s/shared_inp_emb/token_emb/%s"),
("(.*)/(?:decoder|pron_encoder)/token_emb/(.*)",
"%s/shared_out_emb/token_emb/%s")]
# Enable variable sharing.
with py_utils.OpportunisticVariableReuseScope():
with py_utils.VariableRenameScope(renames):
self.CreateChild("encoder", p.encoder)
self.CreateChild("decoder", p.decoder)
if p.use_neighbors:
self.CreateChild("spell_encoder", p.spell_encoder)
if p.pron_encoder:
self.CreateChild("pron_encoder", p.pron_encoder)
def ConstructFPropBPropGraph(self):
# We need to override this since constructing the BPropGraph
# creates slot variables.
p = self._params
with py_utils.OpportunisticVariableReuseScope():
with py_utils.VariableRenameScope(p.variable_renaming_rules):
super(RegExSharedVariableModel,
self).ConstructFPropBPropGraph()
def testRenamingRules(self):
pc = py_utils.WeightParams([3, 3])
with tf.variable_scope('model'):
_, v1 = py_utils.CreateVariable('v1', pc)
with py_utils.VariableRenameScope([('model/(.*)', 'data/%s')]):
_, v2 = py_utils.CreateVariable('v2', pc)
_, v3 = py_utils.CreateVariable('v3', pc)
self.assertTrue(v1.name == 'model/v1/var:0')
self.assertTrue(v2.name == 'data/v2/var:0')
self.assertTrue(v3.name == 'model/v3/var:0')
def __init__(self, train_cfg, ps_params_dict, model_task_name, logdir,
tf_master, **kwargs):
"""Construct an ExecutorTpu BaseRunner.
Args:
train_cfg: SingleTaskModelParams or MultiTaskModelParams
ps_params_dict: A dict of top-level task name -> ProgramSchedule params,
if train_cfg is a SingleTaskModelParams, we expect only one entry.
model_task_name: An override for multi-task models, currently unused.
logdir: String path to the log directory to output to.
tf_master: String path to the master job, e.g. 'local'.
**kwargs: keyword args to pass through to BaseRunner.
"""
super().__init__(train_cfg, model_task_name, logdir, tf_master,
**kwargs)
self._cluster_def = self._cluster.worker_cluster_def
# There is a single Executor task
assert self._cluster.num_replicas == 1
data_parallelism = self._cluster.num_splits_per_client
assert data_parallelism
num_devices_per_split = self._cluster.num_devices_per_split
tf.logging.info('data_parallelism: %d, num_devices_per_split: %d',
data_parallelism, num_devices_per_split)
self.task_scheduler = None
self._checkpoint_dir = os.path.join(logdir, 'train')
self._variable_renaming_rules = []
self._ml_perf = None
# If this is a multi-task model, grab the params for the TaskScheduler.
if issubclass(train_cfg.cls, base_model.SingleTaskModel):
tf.logging.info('single_task_model')
assert len(ps_params_dict) == 1
self._model_task_name = list(ps_params_dict.keys())[0]
self._single_task_mode = True
elif issubclass(train_cfg.cls, base_model.MultiTaskModel):
tf.logging.info('multi_task_model')
if issubclass(train_cfg.cls,
multitask_model.RegExSharedVariableModel):
self._variable_renaming_rules = train_cfg.variable_renaming_rules
if train_cfg.task_schedule is None:
task_schedule_params = task_scheduler.ConstantScheduler.Params(
)
task_schedule_params.task_probs = sorted(
list(train_cfg.task_probs.IterParams()))
else:
task_schedule_params = train_cfg.task_schedule
self.task_scheduler = task_schedule_params.Instantiate()
self._single_task_mode = False
else:
tf.logging.fatal(
'Model %s is not a sub-class of SingleTaskModel or MultiTaskModel',
train_cfg.cls)
tf.logging.info('train_cfg.cls: %s', train_cfg.cls)
self._WriteToLog(train_cfg.ToText(), self._checkpoint_dir,
'trainer_params.txt')
if self._ml_perf is not None:
self._ml_perf_log = True
mlp_log.mlperf_print(key='benchmark',
value=self._ml_perf.benchmark_name)
else:
self._ml_perf_log = False
# BaseRunner legacy
self.enqueue_ops = None
@py_utils.RetryOnTransientTfError()
def _WaitTillInit():
"""Wait until the model is ready."""
try:
with self._graph.as_default(), self._GetSession(
cluster_def=self._cluster_def,
disable_meta_optimizer=FLAGS.
disable_meta_optimizer_in_executor) as sess:
topology = sess.run(
tf.tpu.initialize_system(embedding_config=None,
job=None))
device_assignment = device_assignment_lib.device_assignment(
topology,
computation_shape=py_utils.ComputationShape(
num_devices_per_split, topology),
num_replicas=data_parallelism)
py_utils.SetTpuDeviceAssignment(device_assignment)
tf.logging.info('device_assignment.core_assignment: %s',
str(device_assignment.core_assignment))
tf.logging.info(
'device_assignment.topology.device_coordinates: %s',
str(device_assignment.topology.device_coordinates))
except py_utils.transient_tf_errors as e:
tf.logging.info('TPU initialization failed: %s', e)
raise
if self._ml_perf_log:
mlp_log.mlperf_print(key='init_start', value=None)
_WaitTillInit()
train_cfg = self.params
shared_model = self._MaybeConstructSharedModel(train_cfg)
self._program_schedule_dict = {}
self._programs = []
for task_string, program_schedule_params in ps_params_dict.items():
program_schedule_params.logdir = logdir
program_schedule_params.num_splits_per_client = data_parallelism
program_schedule_params.task_name = task_string
# If the model was created above, we'll inject it here as a shared_model.
ps = program_schedule_params.Instantiate(shared_model=shared_model)
self._program_schedule_dict[task_string] = ps
tf.logging.info('program_schedule_params: %s',
program_schedule_params.ToText())
self._programs += ps.Programs()
if program_schedule_params.ml_perf.benchmark_name is not None:
self._ml_perf = program_schedule_params.ml_perf
tf.logging.info('num_programs: %d', len(self._programs))
with self._graph.as_default(), tf.container(self._container_id):
with self._cluster, tf.device(
self._cluster.job_spec.name if not FLAGS.
cluster_placer_in_executor else self._cluster.GetPlacer()):
with py_utils.VariableRenameScope(
self._variable_renaming_rules):
_ = py_utils.GetOrCreateGlobalStepVar()
for program in self._programs:
program.BuildTpuSubgraph()
py_utils.ClearTpuSummaryTensors()
for program in self._programs:
program.SetStatusMessageFn(self._SetStatusMessage)
program.CreateCheckpointer()
self._initialize_tables = tf.tables_initializer()
self._initialize_local_vars = tf.local_variables_initializer()
self.save_only_checkpointer = checkpointer.Checkpointer(
self._checkpoint_dir,
model=None,
train_params=train_cfg.train,
save_only=True)
def __init__(self, train_cfg, ps_params_dict, model_task_name, logdir,
tf_master, **kwargs):
"""Construct an ExecutorTpu BaseRunner.
Args:
train_cfg: SingleTaskModelParams or MultiTaskModelParams
ps_params_dict: A dict of top-level task name -> ProgramSchedule params,
if train_cfg is a SingleTaskModelParams, we expect only one entry.
model_task_name: An override for multi-task models, currently unused.
logdir: String path to the log directory to output to.
tf_master: String path to the master job, e.g. 'local'.
**kwargs: keyword args to pass through to BaseRunner.
"""
super().__init__(train_cfg, model_task_name, logdir, tf_master,
**kwargs)
data_parallelism = self._cluster.num_splits_per_client
assert data_parallelism
num_devices_per_split = self._cluster.num_devices_per_split
tf.logging.info('data_parallelism: %d, num_devices_per_split: %d',
data_parallelism, num_devices_per_split)
self.task_scheduler = None
self._checkpoint_dir = os.path.join(logdir, 'train')
self._variable_renaming_rules = []
self._ml_perf = None
# If this is a multi-task model, grab the params for the TaskScheduler.
if issubclass(train_cfg.cls, base_model.SingleTaskModel):
tf.logging.info('single_task_model')
assert len(ps_params_dict) == 1
self._model_task_name = list(ps_params_dict.keys())[0]
self._single_task_mode = True
elif issubclass(train_cfg.cls, base_model.MultiTaskModel):
tf.logging.info('multi_task_model')
if issubclass(train_cfg.cls,
multitask_model.RegExSharedVariableModel):
self._variable_renaming_rules = train_cfg.variable_renaming_rules
if train_cfg.task_schedule is None:
task_schedule_params = task_scheduler.ConstantScheduler.Params(
)
task_schedule_params.task_probs = sorted(
list(train_cfg.task_probs.IterParams()))
else:
task_schedule_params = train_cfg.task_schedule
self.task_scheduler = task_schedule_params.Instantiate()
self._single_task_mode = False
else:
tf.logging.fatal(
'Model %s is not a sub-class of SingleTaskModel or MultiTaskModel',
train_cfg.cls)
tf.logging.info('train_cfg.cls: %s', train_cfg.cls)
self._WriteToLog(train_cfg.ToText(), self._checkpoint_dir,
'trainer_params.txt')
if self._ml_perf is not None:
self._ml_perf_log = True
mlp_log.mlperf_print(key='benchmark',
value=self._ml_perf.benchmark_name)
else:
self._ml_perf_log = False
# BaseRunner legacy
self.enqueue_ops = None
train_cfg = self.params
@py_utils.RetryOnTransientTfError()
def _WaitTillInit(job=None):
"""Wait until the model is ready."""
try:
# tpu.initialize_system() is called with None as embedding_config, as
# embedding_config is not available yet. Later in _Loop, it is called
# with the correct embedding_config. Since it cannot be called twice in
# the same graph with different embedding_config, we use a dummy_graph
# here.
dummy_graph = tf.Graph()
with dummy_graph.as_default():
tpu_initialize_system_op = tf.tpu.initialize_system(
embedding_config=None, job=job)
with self._GetSession(graph=dummy_graph) as sess:
topology = sess.run(tpu_initialize_system_op)
if train_cfg.train.tpu_device_order_mode is None:
device_assignment = device_assignment_lib.device_assignment(
topology,
computation_shape=py_utils.ComputationShape(
num_devices_per_split, topology),
num_replicas=data_parallelism)
else:
device_assignment = device_assignment_lib.device_assignment(
topology,
computation_shape=py_utils.ComputationShape(
num_devices_per_split, topology),
num_replicas=data_parallelism,
device_order_mode=train_cfg.train.tpu_device_order_mode
)
py_utils.SetTpuDeviceAssignment(device_assignment, job)
tf.logging.info('device_assignment.core_assignment: %s',
str(device_assignment.core_assignment))
tf.logging.info(
'device_assignment.topology.device_coordinates: %s',
str(device_assignment.topology.device_coordinates))
except py_utils.transient_tf_errors as e:
tf.logging.info('TPU initialization failed: %s', e)
raise
if self._ml_perf_log:
mlp_log.mlperf_print(key='init_start', value=None)
if len(self._cluster.all_worker_names) > 1:
for worker in self._cluster.all_worker_names:
_WaitTillInit(worker)
else:
_WaitTillInit(None)
shared_model = self._MaybeConstructSharedModel(train_cfg)
self._program_schedule_dict = {}
self._programs = []
for task_string, program_schedule_params in ps_params_dict.items():
program_schedule_params.logdir = logdir
program_schedule_params.num_splits_per_client = data_parallelism
program_schedule_params.task_name = task_string
# If the model was created above, we'll inject it here as a shared_model.
ps = program_schedule_params.Instantiate(shared_model=shared_model,
tf_master=self._tf_master)
self._program_schedule_dict[task_string] = ps
tf.logging.info('program_schedule_params: %s',
program_schedule_params.ToText())
self._programs += ps.Programs()
if program_schedule_params.ml_perf.benchmark_name is not None:
self._ml_perf = program_schedule_params.ml_perf
tf.logging.info('num_programs: %d', len(self._programs))
with self._graph.as_default(), tf.container(self._container_id):
with self._cluster, tf.device(self._cluster.GetPlacer()):
with py_utils.VariableRenameScope(
self._variable_renaming_rules):
_ = py_utils.GetOrCreateGlobalStepVar()
for program in self._programs:
program.BuildTpuSubgraph()
py_utils.ClearTpuSummaryTensors()
self._initialize_tables = tf.tables_initializer()
self._initialize_local_vars = tf.local_variables_initializer()
self._initialize_global_vars = tf.global_variables_initializer(
)
for program in self._programs:
program.SetStatusMessageFn(self._SetStatusMessage)
program.CreateCheckpointer(
init_op=self._initialize_global_vars)
self.save_only_checkpointer = checkpointer.Checkpointer(
self._checkpoint_dir,
model=None,
init_op=self._initialize_global_vars,
train_params=train_cfg.train,
save_only=True)
self._load_ops = tf.get_collection(py_utils.TPU_EMBEDDING_LOAD_OPS)
self._retrieve_ops = tf.get_collection(
py_utils.TPU_EMBEDDING_RETRIEVE_OPS)
tpu_embedding_collection = tf.get_collection(
py_utils.TPU_EMBEDDING)
self._tpu_embedding = (tpu_embedding_collection[0]
if tpu_embedding_collection else None)
tf.io.write_graph(self._graph.as_graph_def(), self._checkpoint_dir,
'train.pbtxt')
def __init__(self, train_cfg, ps_params_dict, *args, **kwargs):
"""Construct an ExecutorTpu BaseRunner.
Args:
train_cfg: SingleTaskModelParams or MultiTaskModelParams
ps_params_dict: A dict of top-level task name -> ProgramSchedule params,
if train_cfg is a SingleTaskModelParams, we expect only one entry.
*args: List args to pass through to BaseRunner.
**kwargs: keyword args to pass through to BaseRunner.
"""
if py_utils.IsEagerMode():
assert tf.executing_eagerly()
tf.logging.info(f'FLAGS.tf_master: {FLAGS.tf_master}')
# Connect to the TPU runtime.
resolver = tf.distribute.cluster_resolver.TPUClusterResolver(
FLAGS.tf_master, job_name=FLAGS.worker_job[len('/job:'):])
tf.config.experimental_connect_to_cluster(resolver)
super().__init__(train_cfg, *args, **kwargs)
data_parallelism = self._cluster.num_splits_per_client
assert data_parallelism
num_devices_per_split = self._cluster.num_devices_per_split
tf.logging.info('data_parallelism: %d, num_devices_per_split: %d',
data_parallelism, num_devices_per_split)
self.task_scheduler = None
self._checkpoint_dir = os.path.join(self._logdir, 'train')
self._variable_renaming_rules = []
self._ml_perf = None
# If this is a multi-task model, grab the params for the TaskScheduler.
if issubclass(train_cfg.cls, base_model.SingleTaskModel):
tf.logging.info('single_task_model')
assert len(ps_params_dict) == 1
self._model_task_name = list(ps_params_dict.keys())[0]
self._single_task_mode = True
elif issubclass(train_cfg.cls, base_model.MultiTaskModel):
tf.logging.info('multi_task_model')
if issubclass(train_cfg.cls,
multitask_model.RegExSharedVariableModel):
self._variable_renaming_rules = train_cfg.variable_renaming_rules
if train_cfg.task_schedule is None:
task_schedule_params = task_scheduler.ConstantScheduler.Params(
)
task_schedule_params.task_probs = sorted(
list(train_cfg.task_probs.IterParams()))
else:
task_schedule_params = train_cfg.task_schedule
self.task_scheduler = task_schedule_params.Instantiate()
self._single_task_mode = False
else:
tf.logging.fatal(
'Model %s is not a sub-class of SingleTaskModel or MultiTaskModel',
train_cfg.cls)
tf.logging.info('train_cfg.cls: %s', train_cfg.cls)
self._WriteToLog(train_cfg.ToText(), self._checkpoint_dir,
'trainer_params.txt')
self._WriteToLog(
text_format.MessageToString(train_cfg.ToProto(), as_utf8=True),
self._checkpoint_dir, 'trainer_params.pbtxt')
if self._ml_perf is not None:
self._ml_perf_log = True
mlp_log.mlperf_print(key='benchmark',
value=self._ml_perf.benchmark_name)
else:
self._ml_perf_log = False
train_cfg = self.params
@py_utils.RetryOnTransientTfError()
def _WaitTillInit(job=None):
"""Wait until the model is ready."""
try:
if py_utils.IsEagerMode():
topology = tf.tpu.experimental.initialize_tpu_system(
resolver)
else:
# tpu.initialize_system() is called with None as embedding_config, as
# embedding_config is not available yet. Later in _Loop, it is called
# with the correct embedding_config. Since it cannot be called twice
# in the same graph with different embedding_config, we use a
# dummy_graph here.
dummy_graph = tf.Graph()
with dummy_graph.as_default():
tpu_initialize_system_op = tf.tpu.initialize_system(
embedding_config=None, job=job)
with self._GetSession(graph=dummy_graph) as sess:
topology = sess.run(tpu_initialize_system_op)
if train_cfg.train.tpu_computation_shape is None:
computation_shape = py_utils.ComputationShape(
num_devices_per_split, topology)
else:
computation_shape = train_cfg.train.tpu_computation_shape
assert num_devices_per_split == np.prod(computation_shape)
if train_cfg.train.tpu_device_order_mode is None:
self.device_assignment = device_assignment_lib.device_assignment(
topology,
computation_shape=computation_shape,
num_replicas=data_parallelism)
else:
self.device_assignment = device_assignment_lib.device_assignment(
topology,
computation_shape=computation_shape,
num_replicas=data_parallelism,
device_order_mode=train_cfg.train.tpu_device_order_mode
)
py_utils.SetTpuDeviceAssignment(self.device_assignment, job)
tf.logging.info('device_assignment.core_assignment: %s',
str(self.device_assignment.core_assignment))
tf.logging.info(
'device_assignment.topology.device_coordinates: %s',
str(self.device_assignment.topology.device_coordinates))
except py_utils.transient_tf_errors as e:
tf.logging.info('TPU initialization failed: %s', e)
raise
if self._ml_perf_log:
mlp_log.mlperf_print(key='init_start', value=None)
if len(self._cluster.all_worker_names) > 1:
for worker in self._cluster.all_worker_names:
_WaitTillInit(worker)
else:
_WaitTillInit(None)
shared_model = self._MaybeConstructSharedModel(train_cfg)
self._program_schedule_dict = {}
self._programs = []
self._ckpt_programs = []
self._checkpoint_to_load = None
with self._cluster:
# Create the ExponentialMovingAverage singleton shared by all programs, if
# applicable.
ema = py_utils.CreateEMAForModel(train_cfg, self._global_step_var)
for task_string, program_schedule_params in ps_params_dict.items():
program_schedule_params.logdir = self._logdir
program_schedule_params.num_splits_per_client = data_parallelism
program_schedule_params.task_name = task_string
# If the model was created above, we'll inject it here as a
# shared_model.
ps = program_schedule_params.Instantiate(
shared_model=shared_model,
trial=self._trial,
ema=ema,
tf_master=self._tf_master)
self._program_schedule_dict[task_string] = ps
tf.logging.info('program_schedule_params: %s',
program_schedule_params.ToText())
self._programs += ps.Programs()
if ps.train_program:
self._ckpt_programs.append(ps.train_program)
else:
self._ckpt_programs += ps.Programs()
if program_schedule_params.ml_perf.benchmark_name is not None:
self._ml_perf = program_schedule_params.ml_perf
if ('checkpoint_to_load' in program_schedule_params
and program_schedule_params.checkpoint_to_load):
if (self._checkpoint_to_load
and (self._checkpoint_to_load !=
program_schedule_params.checkpoint_to_load)):
raise ValueError(
f'Multiple values found for checkpoint_to_load: '
f'{self._checkpoint_to_load}, '
f'{program_schedule_params.checkpoint_to_load}.')
self._checkpoint_to_load = program_schedule_params.checkpoint_to_load
tf.logging.info('num_programs: %d', len(self._programs))
# When running in a vizier trainer, the executor reports infeasiable runs
# in case of errors. The programs report metrics and normal completions.
for program in self._programs:
if program._should_report_metrics:
self._should_report_metrics = True
with self._cluster, tf.container(
self._container_id), contextlib.ExitStack() as stack:
if not py_utils.IsEagerMode():
stack.enter_context(self._graph.as_default())
if FLAGS.use_tpu_mirrored_vars:
resolver = tf.distribute.cluster_resolver.TPUClusterResolver(
FLAGS.tf_master,
job_name=FLAGS.worker_job[len('/job:'):])
self._tpu_strategy = tf.distribute.experimental.TPUStrategy(
resolver, device_assignment=self.device_assignment)
stack.enter_context(self._tpu_strategy.scope())
stack.enter_context(
tpu_strategy._TPUReplicaContext(self._tpu_strategy))
else:
stack.enter_context(tf.device(self._cluster.GetPlacer()))
if FLAGS.pdb_on_exception:
stack.enter_context(pdb_wrapper.catch_post_mortem())
with py_utils.VariableStore(), py_utils.VariableRenameScope(
self._variable_renaming_rules):
# `BuildTpuSubgraph` has to be called before checkpoint restore, so that
# the optimizer slot variables are guaranteed to be initialized before
# they get loaded. Otherwise, the optimizers' slot variables will not
# be properly loaded when V1 checkpoint is used.
for program in self._programs:
program.BuildTpuSubgraph()
py_utils.ClearTpuSummaryTensors()
if not py_utils.IsEagerMode():
self._initialize_tables = tf.tables_initializer()
self._initialize_local_vars = tf.local_variables_initializer()
self._initialize_global_vars = tf.global_variables_initializer(
)
checkpointer_models = [
program.GetModel() for program in self._ckpt_programs
]
if py_utils.IsEagerMode():
if FLAGS.use_v2_checkpoints_in_eager:
self._checkpointer = checkpointer.EagerCheckpointerV2(
self._checkpoint_dir,
models=checkpointer_models,
init_op=None,
train_params=train_cfg.train,
save_only=False)
else:
self._checkpointer = checkpointer.EagerCheckpointerV1(
self._checkpoint_dir,
models=checkpointer_models,
init_op=None,
train_params=train_cfg.train,
save_only=False)
else:
self._checkpointer = checkpointer.Checkpointer(
self._checkpoint_dir,
models=checkpointer_models,
init_op=self._initialize_global_vars,
train_params=train_cfg.train,
save_only=False)
for program in self._programs:
program.SetStatusMessageFn(self._SetStatusMessage)
tpu_embedding_collection = (
tpu_embedding_layers.TpuEmbeddingCollection.Get())
self._load_ops = tpu_embedding_collection.load_ops
self._retrieve_ops = tpu_embedding_collection.retrieve_ops
self._tpu_embedding = tpu_embedding_collection.tpu_embedding
def __init__(self, params):
# Enable variable sharing.
p = params
with py_utils.OpportunisticVariableReuseScope():
with py_utils.VariableRenameScope(p.variable_renaming_rules):
super(RegExSharedVariableModel, self).__init__(params)
def __init__(self, train_cfg, ps_params_dict, model_task_name, logdir,
tf_master, **kwargs):
"""Construct an ExecutorTpu BaseRunner.
Args:
train_cfg: SingleTaskModelParams or MultiTaskModelParams
ps_params_dict: A dict of top-level task name -> ProgramSchedule params,
if train_cfg is a SingleTaskModelParams, we expect only one entry.
model_task_name: An override for multi-task models, currently unused.
logdir: String path to the log directory to output to.
tf_master: String path to the master job, e.g. 'local'.
**kwargs: keyword args to pass through to BaseRunner.
"""
super(ExecutorTpu, self).__init__(train_cfg, model_task_name, logdir,
tf_master, **kwargs)
self._cluster_def = self._cluster.worker_cluster_def
# There is a single Executor task
assert self._cluster.num_replicas == 1
data_parallelism = self._cluster.num_splits_per_client
assert data_parallelism
num_devices_per_split = self._cluster.num_devices_per_split
tf.logging.info('data_parallelism: %d, num_devices_per_split: %d',
data_parallelism, num_devices_per_split)
self.task_scheduler = None
self._checkpoint_dir = os.path.join(logdir, 'train')
self._variable_renaming_rules = []
# If this is a multi-task model, grab the params for the TaskScheduler.
if issubclass(train_cfg.cls, base_model.SingleTaskModel):
tf.logging.info('single_task_model')
assert len(ps_params_dict) == 1
self._model_task_name = list(ps_params_dict.keys())[0]
self._single_task_mode = True
elif issubclass(train_cfg.cls, base_model.MultiTaskModel):
tf.logging.info('multi_task_model')
if issubclass(train_cfg.cls,
multitask_model.RegExSharedVariableModel):
self._variable_renaming_rules = train_cfg.variable_renaming_rules
if train_cfg.task_schedule is None:
task_schedule_params = task_scheduler.ConstantScheduler.Params(
)
task_schedule_params.task_probs = sorted(
list(train_cfg.task_probs.IterParams()))
else:
task_schedule_params = train_cfg.task_schedule
self.task_scheduler = task_schedule_params.Instantiate()
self._single_task_mode = False
else:
tf.logging.fatal(
'Model %s is not a sub-class of SingleTaskModel or MultiTaskModel',
train_cfg.cls)
tf.logging.info('train_cfg.cls: %s', train_cfg.cls)
self._WriteToLog(train_cfg.ToText(), self._checkpoint_dir,
'executor_params.txt')
self._program_schedule_dict = {}
self._programs = []
for task_string, program_schedule_params in ps_params_dict.items():
program_schedule_params.logdir = logdir
program_schedule_params.num_splits_per_client = data_parallelism
program_schedule_params.task_name = task_string
ps = program_schedule_params.Instantiate()
self._program_schedule_dict[task_string] = ps
tf.logging.info('program_schedule_params: %s',
program_schedule_params.ToText())
self._programs += ps.Programs()
tf.logging.info('num_programs: %d', len(self._programs))
# BaseRunner legacy
self.enqueue_ops = None
def ComputationShape(split_size):
"""Decides the computation shape based on the split_size."""
computation_shape = None
if split_size == 1:
computation_shape = [1, 1, 1]
elif split_size == 2:
computation_shape = [1, 1, 2]
elif split_size == 4:
computation_shape = [1, 2, 2]
elif split_size == 8:
computation_shape = [2, 2, 2]
elif split_size == 16:
computation_shape = [4, 2, 2]
elif split_size == 32:
computation_shape = [4, 4, 2]
elif split_size == 64:
computation_shape = [4, 8, 2]
elif split_size == 128:
computation_shape = [8, 8, 2]
elif split_size == 256:
computation_shape = [8, 16, 2]
elif split_size == 512:
computation_shape = [16, 16, 2]
else:
assert False, (
'Model parallelism with %d devices is currently not'
' supported.' % split_size)
assert computation_shape is not None
return computation_shape
@py_utils.RetryOnTransientTfError()
def _WaitTillInit():
"""Wait until the model is ready."""
try:
with self._graph.as_default(), self._GetSession(
cluster_def=self._cluster_def) as sess:
topology = sess.run(
tf.tpu.initialize_system(embedding_config=None,
job=None))
device_assignment = device_assignment_lib.device_assignment(
topology,
computation_shape=ComputationShape(
num_devices_per_split),
num_replicas=data_parallelism)
py_utils.SetTpuDeviceAssignment(device_assignment)
tf.logging.info('device_assignment.core_assignment: %s',
str(device_assignment.core_assignment))
tf.logging.info(
'device_assignment.topology.device_coordinates: %s',
str(device_assignment.topology.device_coordinates))
except py_utils.transient_tf_errors as e:
tf.logging.info('TPU initialization failed: %s', e)
raise
_WaitTillInit()
with self._graph.as_default(), tf.container(self._container_id):
with self._cluster, tf.device(
self._cluster.job_spec.name if not FLAGS.
cluster_placer_in_executor else self._cluster.GetPlacer()):
with py_utils.VariableRenameScope(
self._variable_renaming_rules):
for program in self._programs:
program.BuildTpuSubgraph()
for program in self._programs:
program.CreateCheckpointer()
self._initialize_tables = tf.tables_initializer()
self._initialize_local_vars = tf.local_variables_initializer()
self.save_only_checkpointer = checkpointer.Checkpointer(
self._checkpoint_dir,
model=None,
train_params=train_cfg.train,
save_only=True)
def CreateVariables(self):
# Enable variable sharing.
with py_utils.OpportunisticVariableReuseScope():
with py_utils.VariableRenameScope(
self.params.variable_renaming_rules):
super().CreateVariables()
def __init__(self, params):
# Enable variable sharing.
with py_utils.OpportunisticVariableReuseScope():
with py_utils.VariableRenameScope(params.variable_renaming_rules):
super().__init__(params)
