PyTorch torch.ormqr Function
Pytorch torch Reference Manual
torch.ormqrThis is a function in PyTorch that uses QR decomposition for matrix multiplication. It computes Q @ input or input @ Q, where Q is an orthogonal matrix obtained from QR decomposition.
Function Definition
torch.ormqr(input, tau, left=True, transpose=False, out=None)
Parameters:
input(Tensor): Input matrix.tau(Tensor): Householder reflector coefficients.left(bool, optional): If True, computes Q @ input; otherwise computes input @ Q. Defaults to True.transpose(bool, optional): Whether to transpose Q. Defaults to False.out(Tensor, optional): Output tensor.
Return value:
torch.Tensor: Returns the matrix multiplication result.
Usage Example
Example
import torch
# Create matrix
A = torch.randn(3, 3)
B = torch.randn(3, 4)
# QR decomposition
Q, R = torch.linalg.qr(A)
# Compute Q @ B
result = torch.ormqr(B, torch.zeros(3))
print("Shape of matrix B:", B.shape)
print("Result shape:", result.shape)
# Create matrix
A = torch.randn(3, 3)
B = torch.randn(3, 4)
# QR decomposition
Q, R = torch.linalg.qr(A)
# Compute Q @ B
result = torch.ormqr(B, torch.zeros(3))
print("Shape of matrix B:", B.shape)
print("Result shape:", result.shape)
The output result is:
矩阵 B 形状: torch.Size([3, 4]) 结果形状: torch.Size([3, 4])
Other extensions