TopK

public final class TopK

Finds values and indices of the `k` largest elements for the last dimension.

If the input is a vector (rank-1), finds the `k` largest entries in the vector and outputs their values and indices as vectors. Thus `values[j]` is the `j`-th largest entry in `input`, and its index is `indices[j]`.

For matrices (resp. higher rank input), computes the top `k` entries in each row (resp. vector along the last dimension). Thus,

values.shape = indices.shape = input.shape[:-1] + [k]

If two elements are equal, the lower-index element appears first.

Nested Classes

class TopK.Options Optional attributes for TopK  

Constants

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

Public Methods

static <T extends TNumber> TopK<T>
create(Scope scope, Operand<T> input, Operand<TInt32> k, Options... options)
Factory method to create a class wrapping a new TopK operation.
Output<TInt32>
indices()
The indices of `values` within the last dimension of `input`.
static TopK.Options
sorted(Boolean sorted)
Output<T>
values()
The `k` largest elements along each last dimensional slice.

Inherited Methods

Constants

public static final String OP_NAME

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

Constant Value: "TopKV2"

Public Methods

public static TopK<T> create (Scope scope, Operand<T> input, Operand<TInt32> k, Options... options)

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

Parameters
scope current scope
input 1-D or higher with last dimension at least `k`.
k 0-D. Number of top elements to look for along the last dimension (along each row for matrices).
options carries optional attributes values
Returns
  • a new instance of TopK

public Output<TInt32> indices ()

The indices of `values` within the last dimension of `input`.

public static TopK.Options sorted (Boolean sorted)

Parameters
sorted If true the resulting `k` elements will be sorted by the values in descending order.

public Output<T> values ()

The `k` largest elements along each last dimensional slice.