torch_nextafter: Nextafter

In torch: Tensors and Neural Networks with 'GPU' Acceleration

Nextafter

Description

Nextafter

Usage

torch_nextafter(self, other)

Arguments

self	(Tensor) the first input tensor
other	(Tensor) the second input tensor

nextafter(input, other, *, out=None) -> Tensor

Return the next floating-point value after input towards other, elementwise.

The shapes of input and other must be broadcastable .

Examples

if (torch_is_installed()) {

eps <- torch_finfo(torch_float32())$eps
torch_nextafter(torch_tensor(c(1, 2)), torch_tensor(c(2, 1))) == torch_tensor(c(eps + 1, 2 - eps))
}

torch documentation built on June 7, 2023, 6:19 p.m.