credit.losses.base_losses

credit.losses.base_losses#

Attributes#

logger

Functions#

base_losses(conf[, reduction, validation])

Load a specified loss function by its type.

Module Contents#

credit.losses.base_losses.logger#
credit.losses.base_losses.base_losses(conf, reduction='mean', validation=False)#

Load a specified loss function by its type.

Parameters:

  • conf (dict) – Configuration dictionary containing loss settings.

  • reduction (str, optional) – Default reduction method if not specified in parameters.

  • validation (bool) – Use validation loss settings if True, else training loss.

Returns:

Instantiated loss function.

Return type:

torch.nn.Module