def loss_function(y_true, y_pred):
if isinstance(transform, str) and transform.lower() == 'disc':
return losses.discriminative_instance_loss(y_true, y_pred)
if focal:
return losses.weighted_focal_loss(
y_true, y_pred, gamma=gamma, n_classes=n_classes)
return losses.weighted_categorical_crossentropy(
y_true, y_pred, n_classes=n_classes)
def loss_function(y_true, y_pred):
if focal:
return losses.weighted_focal_loss(y_true, y_pred,
gamma=gamma,
n_classes=n_classes,
from_logits=False)
return losses.weighted_categorical_crossentropy(y_true, y_pred,
n_classes=n_classes,
from_logits=False)
def semantic_loss(y_pred, y_true):
return panoptic_weight * losses.weighted_categorical_crossentropy(
y_pred, y_true, n_classes=n_semantic_classes)
def _semantic_loss(y_pred, y_true):
if n_classes > 1:
return panoptic_weight * losses.weighted_categorical_crossentropy(
y_true, y_pred, n_classes=n_classes)
return panoptic_weight * MSE(y_true, y_pred)
