torch_argmax: Argmax

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

Argmax

Description

Argmax

Arguments

self	(Tensor) the input tensor.
dim	(int) the dimension to reduce. If NULL, the argmax of the flattened input is returned.
keepdim	(bool) whether the output tensor has dim retained or not. Ignored if dim=NULL.

argmax(input) -> LongTensor

Returns the indices of the maximum value of all elements in the input tensor.

This is the second value returned by torch_max. See its documentation for the exact semantics of this method.

argmax(input, dim, keepdim=False) -> LongTensor

Returns the indices of the maximum values of a tensor across a dimension.

This is the second value returned by torch_max. See its documentation for the exact semantics of this method.

Examples

if (torch_is_installed()) {

## Not run: 
a = torch_randn(c(4, 4))
a
torch_argmax(a)

## End(Not run)


a = torch_randn(c(4, 4))
a
torch_argmax(a, dim=1)
}

torch documentation built on May 29, 2024, 9:54 a.m.