tensorflow :: ops :: MatrixSetDiagV3

#include <array_ops.h>

Retorna um tensor de matriz em lote com novos valores diagonais em lote.

Resumo

Dado a input e a diagonal , esta operação retorna um tensor com a mesma forma e valores da input , exceto para as diagonais especificadas das matrizes mais internas. Eles serão substituídos pelos valores na diagonal .

input tem r+1 dimensões [I, J, ..., L, M, N] . Quando k é escalar ou k[0] == k[1] , diagonal tem r dimensões [I, J, ..., L, max_diag_len] . Caso contrário, tem r+1 dimensões [I, J, ..., L, num_diags, max_diag_len] . num_diags é o número de diagonais, num_diags = k[1] - k[0] + 1 . max_diag_len é a diagonal mais longa no intervalo [k[0], k[1]] , max_diag_len = min(M + min(k[1], 0), N + min(-k[0], 0))

A saída é um tensor de classificação k+1 com dimensões [I, J, ..., L, M, N] . Se k for escalar ou k[0] == k[1] :

output[i, j, ..., l, m, n]
  = diagonal[i, j, ..., l, n-max(k[1], 0)] ; if n - m == k[1]
    input[i, j, ..., l, m, n]              ; otherwise

Por outro lado,

output[i, j, ..., l, m, n]
  = diagonal[i, j, ..., l, diag_index, index_in_diag] ; if k[0] <= d <= k[1]
    input[i, j, ..., l, m, n]                         ; otherwise
onde d = n - m , diag_index = k[1] - d e index_in_diag = n - max(d, 0) + offset .

offset é zero, exceto quando o alinhamento da diagonal é para a direita.

offset = max_diag_len - diag_len(d) ; if (`align` in {RIGHT_LEFT, RIGHT_RIGHT}
                                           and `d >= 0`) or
                                         (`align` in {LEFT_RIGHT, RIGHT_RIGHT}
                                           and `d <= 0`)
         0                          ; otherwise
onde diag_len(d) = min(cols - max(d, 0), rows + min(d, 0)) .

Por exemplo:

# The main diagonal.
input = np.array([[[7, 7, 7, 7],              # Input shape: (2, 3, 4)
                   [7, 7, 7, 7],
                   [7, 7, 7, 7]],
                  [[7, 7, 7, 7],
                   [7, 7, 7, 7],
                   [7, 7, 7, 7]]])
diagonal = np.array([[1, 2, 3],               # Diagonal shape: (2, 3)
                     [4, 5, 6]])
tf.matrix_set_diag(input, diagonal)
  ==> [[[1, 7, 7, 7],  # Output shape: (2, 3, 4)
        [7, 2, 7, 7],
        [7, 7, 3, 7]],
       [[4, 7, 7, 7],
        [7, 5, 7, 7],
        [7, 7, 6, 7]]]

# A superdiagonal (per batch).
tf.matrix_set_diag(input, diagonal, k = 1)
  ==> [[[7, 1, 7, 7],  # Output shape: (2, 3, 4)
        [7, 7, 2, 7],
        [7, 7, 7, 3]],
       [[7, 4, 7, 7],
        [7, 7, 5, 7],
        [7, 7, 7, 6]]]

# A band of diagonals.
diagonals = np.array([[[0, 9, 1],  # Diagonal shape: (2, 4, 3)
                       [6, 5, 8],
                       [1, 2, 3],
                       [4, 5, 0]],
                      [[0, 1, 2],
                       [5, 6, 4],
                       [6, 1, 2],
                       [3, 4, 0]]])
tf.matrix_set_diag(input, diagonals, k = (-1, 2))
  ==> [[[1, 6, 9, 7],  # Output shape: (2, 3, 4)
        [4, 2, 5, 1],
        [7, 5, 3, 8]],
       [[6, 5, 1, 7],
        [3, 1, 6, 2],
        [7, 4, 2, 4]]]

# LEFT_RIGHT alignment.
diagonals = np.array([[[9, 1, 0],  # Diagonal shape: (2, 4, 3)
                       [6, 5, 8],
                       [1, 2, 3],
                       [0, 4, 5]],
                      [[1, 2, 0],
                       [5, 6, 4],
                       [6, 1, 2],
                       [0, 3, 4]]])
tf.matrix_set_diag(input, diagonals, k = (-1, 2), align="LEFT_RIGHT")
  ==> [[[1, 6, 9, 7],  # Output shape: (2, 3, 4)
        [4, 2, 5, 1],
        [7, 5, 3, 8]],
       [[6, 5, 1, 7],
        [3, 1, 6, 2],
        [7, 4, 2, 4]]]

  

Arguments:

  • scope: A Scope object
  • input: Rank r+1, where r >= 1.
  • diagonal: Rank r when k is an integer or k[0] == k[1]. Otherwise, it has rank r+1. k >= 1.
  • k: Diagonal offset(s). Positive value means superdiagonal, 0 refers to the main diagonal, and negative value means subdiagonals. k can be a single integer (for a single diagonal) or a pair of integers specifying the low and high ends of a matrix band. k[0] must not be larger than k[1].

Optional attributes (see Attrs):

  • align: Some diagonals are shorter than max_diag_len and need to be padded. align is a string specifying how superdiagonals and subdiagonals should be aligned, respectively. There are four possible alignments: "RIGHT_LEFT" (default), "LEFT_RIGHT", "LEFT_LEFT", and "RIGHT_RIGHT". "RIGHT_LEFT" aligns superdiagonals to the right (left-pads the row) and subdiagonals to the left (right-pads the row). It is the packing format LAPACK uses. cuSPARSE uses "LEFT_RIGHT", which is the opposite alignment.

Returns:

  • Output: Rank r+1, with output.shape = input.shape.

Constructors and Destructors

MatrixSetDiagV3(const ::tensorflow::Scope & scope, ::tensorflow::Input input, ::tensorflow::Input diagonal, ::tensorflow::Input k)
MatrixSetDiagV3(const ::tensorflow::Scope & scope, ::tensorflow::Input input, ::tensorflow::Input diagonal, ::tensorflow::Input k, const MatrixSetDiagV3::Attrs & attrs)

Public attributes

operation
output

Public functions

node() const
::tensorflow::Node *
operator::tensorflow::Input() const
operator::tensorflow::Output() const

Public static functions

Align(StringPiece x)

Structs

tensorflow::ops::MatrixSetDiagV3::Attrs

Optional attribute setters for MatrixSetDiagV3.

Public attributes

operation

Operation operation

resultado

::tensorflow::Output output

Funções públicas

MatrixSetDiagV3

 MatrixSetDiagV3(
  const ::tensorflow::Scope & scope,
  ::tensorflow::Input input,
  ::tensorflow::Input diagonal,
  ::tensorflow::Input k
)

MatrixSetDiagV3

 MatrixSetDiagV3(
  const ::tensorflow::Scope & scope,
  ::tensorflow::Input input,
  ::tensorflow::Input diagonal,
  ::tensorflow::Input k,
  const MatrixSetDiagV3::Attrs & attrs
)

::tensorflow::Node * node() const 

operador :: tensorflow :: Input

 operator::tensorflow::Input() const 

operador :: tensorflow :: Saída

 operator::tensorflow::Output() const 

Funções estáticas públicas

Alinhar

Attrs Align(
  StringPiece x
)