BooleanMaskUpdate

public abstract class BooleanMaskUpdate

Nested Classes

class BooleanMaskUpdate.Options Optional attributes for BooleanMaskUpdate  

Public Constructors

BooleanMaskUpdate()

Public Methods

static BooleanMaskUpdate.Options
axis(Integer axis)
Used to indicate the axis to mask from.
static BooleanMaskUpdate.Options
broadcast(Boolean broadcast)
Whether to try broadcasting update.
static <T extends TType> Operand<T>
create(Scope scope, Operand<T> tensor, Operand<TBool> mask, Operand<T> updates, Options... options)
Updates a tensor at the masked values, and returns the updated tensor.

Inherited Methods

Public Constructors

public BooleanMaskUpdate ()

Public Methods

public static BooleanMaskUpdate.Options axis (Integer axis)

Used to indicate the axis to mask from. axis + dim(mask) <= dim(tensor) and mask's shape must match the first axis + dim(mask) dimensions of tensor's shape.

Parameters
axis the axis to mask from. Uses 0 if null.

public static BooleanMaskUpdate.Options broadcast (Boolean broadcast)

Whether to try broadcasting update. True by default.

public static Operand<T> create (Scope scope, Operand<T> tensor, Operand<TBool> mask, Operand<T> updates, Options... options)

Updates a tensor at the masked values, and returns the updated tensor. Does not mutate the input tensors. updates will be broadcasted by default

Numpy equivalent is `tensor[mask] = updates`.

In general, 0 < dim(mask) = K <= dim(tensor), and mask's shape must match the first K dimensions of tensor's shape. We then have: booleanMask(tensor, mask)[i, j1,...,jd] = tensor[i1,...,iK,j1,...,jd] where (i1,...,iK) is the ith true entry of mask (row-major order).

The axis could be used with mask to indicate the axis to mask from (it's 0 by default). In that case, axis + dim(mask) <= dim(tensor) and mask's shape must match the first axis + dim(mask) dimensions of tensor's shape.

The shape of updates should be [n, t_1, t_2, ...] where n is the number of true values in mask and t_i is the ith dimension of tensor after axis and mask. updates will be broadcasted to this shape by default, which can be disabled using options.

Parameters
tensor The tensor to mask.
mask The mask to apply.
updates the new values
options carries optional attributes values
Returns
  • The masked tensor.