Poisson NLL loss

Usage

nn_poisson_nll_loss(
  log_input = TRUE,
  full = FALSE,
  eps = 1e-08,
  reduction = "mean"
)

Arguments

log_input

(bool, optional): if TRUE the loss is computed as $\exp (\text{input})-\text{target}\ast \text{input}$, if FALSE the loss is $\text{input}-\text{target}\ast \log (\text{input}+\text{eps})$.

full

(bool, optional): whether to compute full loss, i. e. to add the Stirling approximation term $\text{target}\ast \log (\text{target})-\text{target}+0.5\ast \log (2\pi \text{target})$.

eps

(float, optional): Small value to avoid evaluation of $\log (0)$ when log_input = FALSE. Default: 1e-8

reduction

(string, optional): Specifies the reduction to apply to the output: 'none' | 'mean' | 'sum'. 'none': no reduction will be applied, 'mean': the sum of the output will be divided by the number of elements in the output, 'sum': the output will be summed.

Details

$$\text{target}\sim \mathrm{Poisson}(\text{input})\text{loss}(\text{input},\text{target})=\text{input}-\text{target}\ast \log (\text{input})+\log (\text{target!})$$

The last term can be omitted or approximated with Stirling formula. The approximation is used for target values more than 1. For targets less or equal to 1 zeros are added to the loss.

Shape

• Input: $(N,\ast )$ where $\ast$ means, any number of additional dimensions

• Target: $(N,\ast )$, same shape as the input

• Output: scalar by default. If reduction is 'none', then $(N,\ast )$, the same shape as the input

Examples

if (torch_is_installed()) {
loss <- nn_poisson_nll_loss()
log_input <- torch_randn(5, 2, requires_grad = TRUE)
target <- torch_randn(5, 2)
output <- loss(log_input, target)
output$backward()
}