Struct auto_diff::op::loss::BCEWithLogitsLoss

pub struct BCEWithLogitsLoss {}

This loss combines a Sigmoid layer and the BCELoss in one single class. This version is more numerically stable than using a plain Sigmoid followed by a BCELoss as, by combining the operations into one layer, we take advantage of the log-sum-exp trick for numerical stability.

-y log (1/(1 + exp(-x))) - (1-y) log(1 - 1/(1 + exp(-x)))

Prediction comes first, label comes second.

Methods

impl BCEWithLogitsLoss

pub fn new() -> BCEWithLogitsLoss

Trait Implementations

impl OpTrait for BCEWithLogitsLoss

fn get_name(&self) -> String

fn apply(&mut self, input: &[&Tensor], output: &[&Tensor])

The first is the prediction, the second input is the label

fn grad(
    &self,
    input: &[&Tensor],
    output_grad: &[&Tensor],
    input_grad: &[&Tensor]
)

Given the forward input value and backward output_grad, Update weight gradient. return backward input gradeint.

fn get_values(&self) -> Vec<&Tensor>

access weight values

fn set_values(&self, v: &[Tensor])

fn get_grads(&self) -> Vec<&Tensor>

access gradient values