nn_threshold function

Threshold module

Thresholds each element of the input Tensor.

nn_threshold(threshold, value, inplace = FALSE)

Arguments

• threshold: The value to threshold at
• value: The value to replace with
• inplace: can optionally do the operation in-place. Default: FALSE

Details

Threshold is defined as:

Shape

• Input: $(N, *)$ where * means, any number of additional dimensions
• Output: $(N, *)$, same shape as the input

Examples

if (torch_is_installed()) {
m <- nn_threshold(0.1, 20)
input <- torch_randn(2)
output <- m(input)
}