nntoolbox.losses.pinball module

class nntoolbox.losses.pinball.PinballLoss(tau: float = 0.5, reduction: str = 'mean')

Bases: torch.nn.modules.module.Module

Pinball loss for quantile regression:

L_tau(y_true, y_pred) = max(tau * (y_true - y_pred), (tau - 1) * (y_true - y_pred))

References:

https://www.tensorflow.org/addons/api_docs/python/tfa/losses/PinballLoss

Ingo Steinwart and Andreas Christmann, “Estimating conditional quantiles with the help of the pinball loss.” https://projecteuclid.org/download/pdfview_1/euclid.bj/1297173840

forward(input: torch.Tensor, target: torch.Tensor) → torch.Tensor

Defines the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

training: bool