เทนเซอร์โฟลว์:: ปฏิบัติการ:: MatrixDiag

#include <array_ops.h>

ส่งกลับเทนเซอร์แนวทแยงแบบแบทช์พร้อมค่าแนวทแยงแบบแบทช์ที่กำหนด

สรุป

เมื่อกำหนด diagonal การดำเนินการนี้จะส่งคืนเทนเซอร์ที่มี diagonal และทุกอย่างเสริมด้วยศูนย์ เส้นทแยงมุมคำนวณดังนี้:

สมมติว่า diagonal มี k มิติ [I, J, K, ..., N] ดังนั้นเอาต์พุตจะเป็นเมตริกซ์ของอันดับ k+1 ด้วยขนาด [I, J, K, ..., N, N]` โดยที่:

output[i, j, k, ..., m, n] = 1{m=n} * diagonal[i, j, k, ..., n]

ตัวอย่างเช่น:

# 'diagonal' is [[1, 2, 3, 4], [5, 6, 7, 8]]

and diagonal.shape = (2, 4)

tf.matrix_diag(diagonal) ==> [[[1, 0, 0, 0]
                                     [0, 2, 0, 0]
                                     [0, 0, 3, 0]
                                     [0, 0, 0, 4]],
                                    [[5, 0, 0, 0]
                                     [0, 6, 0, 0]
                                     [0, 0, 7, 0]
                                     [0, 0, 0, 8]]]

which has shape (2, 4, 4)

ข้อโต้แย้ง:

  • ขอบเขต: วัตถุ ขอบเขต
  • เส้นทแยงมุม: อันดับ k โดยที่ k >= 1

ผลตอบแทน:

  • Output : อันดับ k+1 โดยมี output.shape = diagonal.shape + [diagonal.shape[-1]]

ตัวสร้างและผู้ทำลาย

MatrixDiag (const :: tensorflow::Scope & scope, :: tensorflow::Input diagonal)

คุณลักษณะสาธารณะ

operation
output

งานสาธารณะ

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

คุณลักษณะสาธารณะ

การดำเนินการ

Operation operation

เอาท์พุท

::tensorflow::Output output

งานสาธารณะ

MatrixDiag

 MatrixDiag(
  const ::tensorflow::Scope & scope,
  ::tensorflow::Input diagonal
)

โหนด

::tensorflow::Node * node() const 

ตัวดำเนินการ::tensorflow::อินพุต

 operator::tensorflow::Input() const 

ตัวดำเนินการ::tensorflow::เอาต์พุต

 operator::tensorflow::Output() const