tf.distribute.cluster_resolver.TFConfigClusterResolver

Implementation of a ClusterResolver which reads the TF_CONFIG EnvVar.

Inherits From: ClusterResolver

tf.distribute.cluster_resolver.TFConfigClusterResolver(
    task_type=None, task_id=None, rpc_layer=None, environment=None
)

This is an implementation of cluster resolvers when using TF_CONFIG to set information about the cluster. The cluster spec returned will be initialized from the TF_CONFIG environment variable.

An example to set TF_CONFIG is:

  os.environ['TF_CONFIG'] = json.dumps({
    'cluster': {
        'worker': ["localhost:12345", "localhost:23456"]
    },
    'task': {'type': 'worker', 'index': 0}
  })

However, sometimes the container orchestration framework will set TF_CONFIG for you. In this case, you can just create an instance without passing in any arguments. You can find an example here to let Kuburnetes set TF_CONFIG for you: https://github.com/tensorflow/ecosystem/tree/master/kubernetes. Then you can use it with tf.distribute.Strategy as:

  # `TFConfigClusterResolver` is already the default one in the following
  # strategy.
  strategy = tf.distribute.experimental.MultiWorkerMirroredStrategy(
      cluster_resolver=TFConfigClusterResolver())

task_type (String, optional) Overrides the task type specified in the TF_CONFIG environment variable.
task_id (Integer, optional) Overrides the task index specified in the TF_CONFIG environment variable.
rpc_layer (String, optional) Overrides the rpc layer TensorFlow uses.
environment (String, optional) Overrides the environment TensorFlow operates in.

environment Returns the current environment which TensorFlow is running in.

There are two possible return values, "google" (when TensorFlow is running in a Google-internal environment) or an empty string (when TensorFlow is running elsewhere).

If you are implementing a ClusterResolver that works in both the Google environment and the open-source world (for instance, a TPU ClusterResolver or similar), you will have to return the appropriate string depending on the environment, which you will have to detect.

Otherwise, if you are implementing a ClusterResolver that will only work in open-source TensorFlow, you do not need to implement this property.

rpc_layer

task_id Returns the task id this ClusterResolver indicates.

In TensorFlow distributed environment, each job may have an applicable task id, which is the index of the instance within its task type. This is useful when user needs to run specific code according to task index. For example,

cluster_spec = tf.train.ClusterSpec({
"ps": ["localhost:2222", "localhost:2223"],
"worker": ["localhost:2224", "localhost:2225", "localhost:2226"]
})

# SimpleClusterResolver is used here for illustration; other cluster
# resolvers may be used for other source of task type/id.
simple_resolver = SimpleClusterResolver(cluster_spec, task_type="worker",
task_id=0)

...

if cluster_resolver.task_type == 'worker' and cluster_resolver.task_id == 0:
# Perform something that's only applicable on 'worker' type, id 0. This
# block will run on this particular instance since we've specified this
# task to be a 'worker', id 0 in above cluster resolver.
else:
# Perform something that's only applicable on other ids. This block will
# not run on this particular instance.

Returns None if such information is not available or is not applicable in the current distributed environment, such as training with tf.distribute.cluster_resolver.TPUClusterResolver.

For more information, please see tf.distribute.cluster_resolver.ClusterResolver's class docstring.

task_type Returns the task type this ClusterResolver indicates.

In TensorFlow distributed environment, each job may have an applicable task type. Valid task types in TensorFlow include 'chief': a worker that is designated with more responsibility, 'worker': a regular worker for training/evaluation, 'ps': a parameter server, or 'evaluator': an evaluator that evaluates the checkpoints for metrics.

See Multi-worker configuration for more information about 'chief' and 'worker' task type, which are most commonly used.

Having access to such information is useful when user needs to run specific code according to task types. For example,

cluster_spec = tf.train.ClusterSpec({
"ps": ["localhost:2222", "localhost:2223"],
"worker": ["localhost:2224", "localhost:2225", "localhost:2226"]
})

# SimpleClusterResolver is used here for illustration; other cluster
# resolvers may be used for other source of task type/id.
simple_resolver = SimpleClusterResolver(cluster_spec, task_type="worker",
task_id=1)

...

if cluster_resolver.task_type == 'worker':
# Perform something that's only applicable on workers. This block
# will run on this particular instance since we've specified this task to
# be a worker in above cluster resolver.
elif cluster_resolver.task_type == 'ps':
# Perform something that's only applicable on parameter servers. This
# block will not run on this particular instance.

Returns None if such information is not available or is not applicable in the current distributed environment, such as training with tf.distribute.experimental.TPUStrategy.

For more information, please see tf.distribute.cluster_resolver.ClusterResolver's class doc.

Methods

cluster_spec

View source

cluster_spec()

Returns a ClusterSpec based on the TF_CONFIG environment variable.

Returns
A ClusterSpec with information from the TF_CONFIG environment variable.

master

View source

master(
    task_type=None, task_id=None, rpc_layer=None
)

Returns the master address to use when creating a TensorFlow session.

Args
task_type (String, optional) Overrides and sets the task_type of the master.
task_id (Integer, optional) Overrides and sets the task id of the master.
rpc_layer (String, optional) Overrides and sets the protocol over which TensorFlow nodes communicate with each other.

Returns
The address of the master.

Raises
RuntimeError If the task_type or task_id is not specified and the TF_CONFIG environment variable does not contain a task section.

num_accelerators

View source

num_accelerators(
    task_type=None, task_id=None, config_proto=None
)

Returns the number of accelerator cores per worker.

This returns the number of accelerator cores (such as GPUs and TPUs) available per worker.

Optionally, we allow callers to specify the task_type, and task_id, for if they want to target a specific TensorFlow task to query the number of accelerators. This is to support heterogenous environments, where the number of accelerators cores per host is different.

Args
task_type (Optional) The type of the TensorFlow task of the machine we want to query.
task_id (Optional) The index of the TensorFlow task of the machine we want to query.
config_proto (Optional) Configuration for starting a new session to query how many accelerator cores it has.

Returns
A map of accelerator types to number of cores.