Convenience wrapper to create Numpy arrays with a np.float32 dtype.

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

Return a slot named `name` created for `var` by the Optimizer. Some `Optimizer` subclasses use additional variables. For example `Momentum`

The vdotq_lane_s32 takes int8x8t for the rhs parameter, whereas the actual instruction selects from between 4 32-bit (4x8-bit packed) sub-registers, a

The device for SaveSpec Tensors.

Add operations to minimize `loss` by updating `var_list`. This method simply combines calls `compute_gradients()` and `apply_gradients()`. If

Returns index of node in GraphDef/GraphView.

Probability density/mass function. Args: value: `float` or `double` `Tensor`. name: Python `str` prepended to names of ops created by

Return whether the rank is unknown

Converts a TensorFlow GraphDef based on instance variables. Returns: The converted data in serialized format. Either a TFLite Flatbuffer or

Returns a NumPy array specifying the minimum bounds (inclusive).

Only needed by variable op to set the shape of an uninitialized Tensor. TODO: Remove this when we have a better story for detecting uninitialized tens

Returns the size of the file.

Run the test, and return 0 on success, 2 otherwise.

`high - low`.

Releases the cleanup function instead of running it. Hint: use c.release()() to run early.

Whether this buffer contains an array.

Returns NodeView from `graph_view_` at `node_index_`.

Resets the RefPtr from a pointer. This is an ownership transfer operation, and the caller has to own a valid reference to ptr (unless ptr is nullptr).

Note that it is not required that the first dimension be the number of batches. But it is required that the last dimension is the 'input_dim'. For th

Returns the data type.

Returns the result of the Metric. Args: write_summary: bool indicating whether to feed the result to the summary before returning.

Fit the estimator using the training data. Train the estimator for `self._train_steps` steps, after waiting for `delay_secs` seconds. If `sel

Returns the total count of elements, that is the size when flattened into a vector.

Verifies the shape or the shape and the layout of an HLO instruction against the provided shape object.

Returns a context manager that enables summary writing.

A deterministic truncated normal if seed is passed.

Matches a node with op `op`.

Adds a new variable to the layer. Arguments: name: Variable name. shape: Variable shape. Defaults to scalar if unspecified. dty

Shape of a single sample from a single batch as a 1-D int32 `Tensor`. Args: name: name to give to the op Returns: event_shape: `

If you know how many elements you will push at the time you create the TopN object, you can call reserve to preallocate the memory that TopN will need

Make and run an experiment. It creates an Experiment by calling `experiment_fn`. Then it calls the function named as `schedule` of the Experiment

Sequence of variables owned by this module and it's submodules. Note: this method uses reflection to find variables on the current instance a

Returns the number of users of this instruction.

The device of this variable.

Store export information for constants/string literals. Export information is stored in the module where constants/string literals are define

Call `metric_value_fn` in the correct control flow context.

Returns a `RaggedTensor` containing the specified sequences of numbers. Each row of the returned `RaggedTensor` contains a single sequence: ```p

Returns a new `tf.data.Dataset` with the given options set. The options are "global" in the sense they apply to the entire dataset. If option

static */