tf_quant_finance.utils.broadcast_common_batch_shape

Last updated: 2023-03-16.

tf_quant_finance.utils.broadcast_common_batch_shape#

View source

Broadcasts argument batch shapes to the common shape.

tf_quant_finance.utils.broadcast_common_batch_shape(
    *, event_ranks=None, name=None, *args
)

Each input Tensor is assumed to be of shape batch_shape_i + event_shape_i. The function finds a common batch_shape and broadcasts each Tensor to batch_shape + event_shape_i. The common batch shape is the minimal shape such that all batch_shape_i can broadcast to it.

Example 1. Batch shape is all dimensions but the last one#

import tensorflow as tf
import tf_quant_finance as tff

# Two Tensors of shapes [2, 3] and [2]. The batch shape of the 1st Tensor is
# [2] and for the second is []. The common batch shape is [2]
args = [tf.ones([2, 3], dtype=tf.float64), tf.constant([True, False])]
tff.utils.broadcast_common_batch_shape(*args)
# Expected: (array([[1., 1., 1.], [1., 1., 1.]]),
#            array([[True, True], [False, False]])

Example 2. Specify ranks of event shapes#

import tensorflow as tf
import tf_quant_finance as tff

args = [tf.ones([2, 3], dtype=tf.float64), tf.constant([True, False])]
tff.utils.broadcast_common_batch_shape(*args,
                                       event_ranks)
# Expected: (array([[1., 1., 1.], [1., 1., 1.]]),
#            array([[True, True], [False, False]])

Args:#

  • *args: A sequence of Tensors of compatible shapes and any dtypes.

  • event_ranks: A sequence of integers of the same length as args specifying ranks of event_shape for each input Tensor. Default value: None which means that all dimensions but the last one are treated as batch dimension.

  • name: Python string. The name to give to the ops created by this function. Default value: None which maps to the default name broadcast_tensor_shapes.

Returns:#

A tuple of broadcasted Tensors. Each Tensor has the same dtype as the corresponding input Tensor.

Raises:#

  • ValueError: (a) If event_ranks is supplied and is of different from args length. (b) If inputs are of incompatible shapes.

Contents