Dequeues and concatenates `n` elements from this queue. This operation concatenates queue-element component tensors along the 0th dimension t

Trains a model given training data `x` predictions and `y` labels. Args: x: Matrix of shape [n_samples, n_features...] or the dictionary of

-- ACCESSORS -------------------------------------------------------------- The class provides a value() accessor returning the stored ValueType value

Returns the buffer defined at the given instruction and index.

\brief Copy the other tensor into this tensor and reshape it. This tensor shares other's underlying storage. Returns `true` iff `other.shape()` has t

Asserts that ProtoEq says a != b.

Returns `ModelFnOps` that a model_fn can return. Please note that, + Exactly one of `logits` and `logits_input` must be provided. + All a

\brief data() points to a memory region of size() bytes. NOTE(mrry): The `data()` method is not virtual for performance reasons. It can be called mul

r"""Gather slices from `params` into a Tensor with shape specified by `indices`. See tf.gather_nd for details. Args: indices: A `Tenso

Return the unique identifier for this HloBuffer.

Blocks until the thread terminates. Raises: self._testcase.failureException: If the thread terminates with due to an exceptio

Returns a new dimension of the given size. The returned value is owned by this context.

Raises an exception unless a non-trivial restoration has completed.

Allocate the second output tensor that will contain the MKL shape serialized

Parse the following regex manually name(\[index\])?(\.field)?

Apply gradients to variables. This is the second part of `minimize()`. It returns an `Operation` that applies gradients. Args: gra

Creates an `Iterator` for enumerating the elements of this dataset. Note: The returned iterator will be in an uninitialized state, and you mu

Creates an `Iterator` for enumerating the elements of this dataset. Note: The returned iterator will be initialized automatically. A "one-sho

Op converters (e.g. ConvertReshape) need to access the TRT network in order to add TRT layers.

Removes symbols in a module that are not referenced by a docstring. Args: module_name: the name of the module (usually `__name__`). allowed

Restores previously saved variables. This method runs the ops added by the constructor for restoring variables. It requires a session in whic

UpdateErr is used to notify Scope of any graph construction errors while creating the operation op.

Returns word's id in the vocabulary. If category is new, creates a new id for it. Args: category: string or integer to lookup in vocab

Returns the identifier of underlying Servable.

Add a sequence of trackable values.

Returns an id that is unique to this module across all modules created over the lifetime of this process.

Apply gradients to a replica variable.

Flat size calculation, checking that dimensions match with one or more other arrays.

Asserts that two float arrays are near each other. Checks that for all elements of farray1 and farray2 |f1 - f2| < err. Asserts a test failu

Creates a Tensor from a Java object. <p>A {@code Tensor} is a multi-dimensional array of elements of a limited set of types. Not all Java objects can

Find dumped tensor data by a certain predicate. Args: predicate: A callable that takes two input arguments: ```python def

Returns the list of global variables created by the Template.

Writes a `SavedModel` protocol buffer to disk. The function writes the SavedModel protocol buffer to the export directory in serialized forma

Overload to call And with 3 or more operands. We need the following somewhat convoluted overload set to disambiguate with the overload that takes the

Wait for threads to terminate. This call blocks until a set of threads have terminated. The set of thread is the union of the threads passed

Matches nested structures. Recursively matches shape and values of `expected` and `actual`. Handles scalars, numpy arrays and other python se

Returns the number of inputs to the operator.