QR decomposition

Computes the QR decomposition of a matrix input, and returns matrices Q and R such that $input = QR$, with QQ being an orthogonal matrix and RR being an upper triangular matrix.

tch_qr(x)

Arguments

x	tensor object

Details

This returns the thin (reduced) QR factorization.

Examples

x <- tensor(matrix(runif(16), ncol = 4))
tch_qr(x)
#> [[1]]
#> tensor 
#> -0.1476  0.7844 -0.5595 -0.2232
#> -0.7923  0.0812  0.4731 -0.3767
#> -0.5265 -0.5149 -0.6532  0.1760
#> -0.2708  0.3361  0.1909  0.8816
#> [ Variable[CPUFloatType]{4,4} ]
#> 
#> [[2]]
#> tensor 
#> -1.1657 -1.0536 -0.6330 -0.6298
#>  0.0000  0.8340  0.2082 -0.0984
#>  0.0000  0.0000  0.2308 -0.5939
#>  0.0000  0.0000  0.0000  0.1678
#> [ Variable[CPUFloatType]{4,4} ]
#>