Pix2PixLossSemantic

Inheritance Diagram

---

class ashpy.losses.gan.Pix2PixLossSemantic(cross_entropy_weight=100.0, adversarial_loss_weight=1.0, feature_matching_weight=10.0, adversarial_loss_type=<AdversarialLossType.GAN: 1>, use_feature_matching_loss=False)

Bases: ashpy.losses.executor.SumExecutor

Semantic Pix2Pix Loss.

Weighted sum of ashpy.losses.gan.CategoricalCrossEntropy, ashpy.losses.gan.AdversarialLossG and ashpy.losses.gan.FeatureMatchingLoss.

Methods

__init__([cross_entropy_weight, …])

Initialize the Executor.

Attributes

executors	Return the List of Executors.
fn	Return the Keras loss function to execute.
global_batch_size	Global batch size comprises the batch size for each cpu.
weight	Return the loss weight.

__init__(cross_entropy_weight=100.0, adversarial_loss_weight=1.0, feature_matching_weight=10.0, adversarial_loss_type=<AdversarialLossType.GAN: 1>, use_feature_matching_loss=False)

Initialize the Executor.

Weighted sum of ashpy.losses.gan.CategoricalCrossEntropy, ashpy.losses.gan.AdversarialLossG and ashpy.losses.gan.FeatureMatchingLoss

Parameters:

• cross_entropy_weight (ashpy.ashtypes.TWeight) – Weight of the categorical cross entropy loss.
• adversarial_loss_weight (ashpy.ashtypes.TWeight) – Weight of the adversarial loss.
• feature_matching_weight (ashpy.ashtypes.TWeight) – Weight of the feature matching loss.
• adversarial_loss_type (ashpy.losses.gan.AdversarialLossType) – type of adversarial loss, see ashpy.losses.gan.AdversarialLossType
• use_feature_matching_loss (bool) – whether to use feature matching loss or not