tfdbg.TensorBoardDebugWrapperSession

Class TensorBoardDebugWrapperSession

Inherits From: GrpcDebugWrapperSession

Defined in tensorflow/python/debug/wrappers/grpc_wrapper.py.

A tfdbg Session wrapper that can be used with TensorBoard Debugger Plugin.

This wrapper is the same as GrpcDebugWrapperSession, except that it uses a predefined watch_fn that 1) uses DebugIdentity debug ops with the gated_grpc attribute set to True to allow the interactive enabling and disabling of tensor breakpoints. 2) watches all tensors in the graph. This saves the need for the user to define a watch_fn.

Properties

graph

The underlying TensorFlow graph, to be used in building Operations.

graph_def

run_call_count

sess_str

The TensorFlow process to which this session will connect.

session

Methods

__init__

__init__(
    sess,
    grpc_debug_server_addresses,
    thread_name_filter=None,
    send_traceback_and_source_code=True,
    log_usage=True
)

Constructor of TensorBoardDebugWrapperSession.

Args:

• sess: The tf.Session instance to be wrapped.
• grpc_debug_server_addresses: gRPC address(es) of debug server(s), as a str or a list of strs. E.g., "localhost:2333", "grpc://localhost:2333", ["192.168.0.7:2333", "192.168.0.8:2333"].
• thread_name_filter: Optional filter for thread names.
• send_traceback_and_source_code: Whether traceback of graph elements and the source code are to be sent to the debug server(s).
• log_usage: Whether the usage of this class is to be logged (if applicable).

__enter__

__enter__()

__exit__

__exit__(
    exec_type,
    exec_value,
    exec_tb
)

as_default

as_default()

close

close()

increment_run_call_count

increment_run_call_count()

invoke_node_stepper

invoke_node_stepper(
    node_stepper,
    restore_variable_values_on_exit=True
)

See doc of BaseDebugWrapperSession.invoke_node_stepper.

list_devices

list_devices(
    *args,
    **kwargs
)

make_callable

make_callable(
    fetches,
    feed_list=None,
    accept_options=False
)

on_run_end

on_run_end(request)

See doc of BaseDebugWrapperSession.on_run_end.

on_run_start

on_run_start(request)

See doc of BaseDebugWrapperSession.on_run_start.

on_session_init

on_session_init(request)

See doc of BaseDebugWrapperSession.on_run_start.

partial_run

partial_run(
    handle,
    fetches,
    feed_dict=None
)

Continues the execution with additional feeds and fetches.

partial_run_setup

partial_run_setup(
    fetches,
    feeds=None
)

Sets up the feeds and fetches for partial runs in the session.

prepare_run_debug_urls

prepare_run_debug_urls(
    fetches,
    feed_dict
)

Implementation of abstract method in superclass.

See doc of NonInteractiveDebugWrapperSession.prepare_run_debug_urls() for details.

Args:

• fetches: Same as the fetches argument to Session.run()
• feed_dict: Same as the feed_dict argument to Session.run()

Returns:

• debug_urls: (str or list of str) file:// debug URLs to be used in this Session.run() call.

reset

reset(
    *args,
    **kwargs
)

run

run(
    fetches,
    feed_dict=None,
    options=None,
    run_metadata=None,
    callable_runner=None,
    callable_runner_args=None
)

Wrapper around Session.run() that inserts tensor watch options.

Args:

• fetches: Same as the fetches arg to regular Session.run().
• feed_dict: Same as the feed_dict arg to regular Session.run().
• options: Same as the options arg to regular Session.run().
• run_metadata: Same as the run_metadata arg to regular Session.run().
• callable_runner: A callable returned by Session.make_callable(). If not None, fetches and feed_dict must both be None.
• callable_runner_args: An optional list of arguments to callable_runner.

Returns:

Simply forwards the output of the wrapped Session.run() call.

Raises:

• ValueError: On invalid OnRunStartAction value. Or if callable_runner is not None and either or both of fetches and feed_dict is None.

run_step_fn

run_step_fn(step_fn)

should_stop

should_stop()