CumulativeLogsumexp

public final class CumulativeLogsumexp

Compute the cumulative product of the tensor `x` along `axis`.

By default, this op performs an inclusive cumulative log-sum-exp, which means that the first element of the input is identical to the first element of the output:

tf.math.cumulative_logsumexp([a, b, c])  # => [a, log(exp(a) + exp(b)), log(exp(a) + exp(b) + exp(c))]
 
By setting the `exclusive` kwarg to `True`, an exclusive cumulative log-sum-exp is performed instead:
tf.cumulative_logsumexp([a, b, c], exclusive=True)  # => [-inf, a, log(exp(a) * exp(b))]
 
Note that the neutral element of the log-sum-exp operation is `-inf`, however, for performance reasons, the minimal value representable by the floating point type is used instead.

By setting the `reverse` kwarg to `True`, the cumulative log-sum-exp is performed in the opposite direction.

Nested Classes

class CumulativeLogsumexp.Options Optional attributes for CumulativeLogsumexp

Constants

String OP_NAME The name of this op, as known by TensorFlow core engine

Public Methods

Output <T>
asOutput ()
Returns the symbolic handle of the tensor.
static <T extends TNumber > CumulativeLogsumexp <T>
create ( Scope scope, Operand <T> x, Operand <? extends TNumber > axis, Options... options)
Factory method to create a class wrapping a new CumulativeLogsumexp operation.
static CumulativeLogsumexp.Options
exclusive (Boolean exclusive)
Output <T>
static CumulativeLogsumexp.Options
reverse (Boolean reverse)

Inherited Methods

Constants

public static final String OP_NAME

The name of this op, as known by TensorFlow core engine

Constant Value: "CumulativeLogsumexp"

Public Methods

public Output <T> asOutput ()

Returns the symbolic handle of the tensor.

Inputs to TensorFlow operations are outputs of another TensorFlow operation. This method is used to obtain a symbolic handle that represents the computation of the input.

public static CumulativeLogsumexp <T> create ( Scope scope, Operand <T> x, Operand <? extends TNumber > axis, Options... options)

Factory method to create a class wrapping a new CumulativeLogsumexp operation.

Parameters
scope current scope
x A `Tensor`. Must be one of the following types: `float16`, `float32`, `float64`.
axis A `Tensor` of type `int32` (default: 0). Must be in the range `[-rank(x), rank(x))`.
options carries optional attributes values
Returns
  • a new instance of CumulativeLogsumexp

public static CumulativeLogsumexp.Options exclusive (Boolean exclusive)

Parameters
exclusive If `True`, perform exclusive cumulative log-sum-exp.

public Output <T> out ()

public static CumulativeLogsumexp.Options reverse (Boolean reverse)

Parameters
reverse A `bool` (default: False).