log

Returns a new tensor with the logarithm of the elements of input. The available logarithm functions are:

tch_log: Natural logarithm.

tch_log2: Base 2 logarithm.

tch_log10: Base 10 logarithm.

tch_log1p: Same as tch_log(1 + x). This function is more accurate than tch_log() for small values of x.

tch_log(x)

tch_log2(x)

tch_log10(x)

tch_log1p(x)

Arguments

x	tensor object

Examples

x <- tch_randn(c(2,2))

tch_log(x)
#> tensor 
#>     nan -1.2291
#>  0.0796     nan
#> [ Variable[CPUFloatType]{2,2} ]

tch_log2(x)
#> tensor 
#>     nan -1.7732
#>  0.1148     nan
#> [ Variable[CPUFloatType]{2,2} ]

tch_log10(x)
#> tensor 
#>     nan -0.5338
#>  0.0346     nan
#> [ Variable[CPUFloatType]{2,2} ]

tch_log1p(x)
#> tensor 
#>     nan  0.2566
#>  0.7337     nan
#> [ Variable[CPUFloatType]{2,2} ]