gan

GANContext measures the specified metrics on the GAN.

Classes

GANContext

ashpy.contexts.gan.GANContext measure the specified metrics on the GAN.

GANEncoderContext

ashpy.contexts.gan.GANEncoderContext measure the specified metrics on the GAN.
class ashpy.contexts.gan.GANContext(dataset=None, generator_model=None, discriminator_model=None, generator_loss=None, discriminator_loss=None, metrics=None, log_eval_mode=<LogEvalMode.TRAIN: 1>, global_step=<tf.Variable 'global_step:0' shape=() dtype=int64, numpy=0>, ckpt=None)[source]

Bases: ashpy.contexts.base_context.BaseContext

ashpy.contexts.gan.GANContext measure the specified metrics on the GAN.

__init__(dataset=None, generator_model=None, discriminator_model=None, generator_loss=None, discriminator_loss=None, metrics=None, log_eval_mode=<LogEvalMode.TRAIN: 1>, global_step=<tf.Variable 'global_step:0' shape=() dtype=int64, numpy=0>, ckpt=None)[source]

Initialize the Context.

Parameters

  • dataset (tf.data.Dataset) – Dataset of tuples. [0] true dataset, [1] generator input dataset.

  • generator_model (tf.keras.Model) – The generator.

  • discriminator_model (tf.keras.Model) – The discriminator.

  • generator_loss (ashpy.losses.Executor()) – The generator loss.

  • discriminator_loss (ashpy.losses.Executor()) – The discriminator loss.

  • metrics (list of [ashpy.metrics.metric.Metric]) – All the metrics to be used to evaluate the model.

  • log_eval_mode (ashpy.modes.LogEvalMode) – Models’ mode to use when evaluating and logging.

  • global_step (tf.Variable) – tf.Variable that keeps track of the training steps.

  • ckpt (tf.train.Checkpoint) – checkpoint to use to keep track of models status.

Return type

None

property discriminator_loss

Retrieve the discriminator loss.

Return type

Optional[Executor]

property discriminator_model

Retrieve the discriminator model.

Return type

Model

Returns

tf.keras.Model.

property generator_loss

Retrieve the generator loss.

Return type

Optional[Executor]

property generator_model

Retrieve the generator model.

Return type

Model

Returns

tf.keras.Model.

class ashpy.contexts.gan.GANEncoderContext(dataset=None, generator_model=None, discriminator_model=None, encoder_model=None, generator_loss=None, discriminator_loss=None, encoder_loss=None, metrics=None, log_eval_mode=<LogEvalMode.TRAIN: 1>, global_step=<tf.Variable 'global_step:0' shape=() dtype=int64, numpy=0>, ckpt=None)[source]

Bases: ashpy.contexts.gan.GANContext

ashpy.contexts.gan.GANEncoderContext measure the specified metrics on the GAN.

__init__(dataset=None, generator_model=None, discriminator_model=None, encoder_model=None, generator_loss=None, discriminator_loss=None, encoder_loss=None, metrics=None, log_eval_mode=<LogEvalMode.TRAIN: 1>, global_step=<tf.Variable 'global_step:0' shape=() dtype=int64, numpy=0>, ckpt=None)[source]

Initialize the Context.

Parameters

  • dataset (tf.data.Dataset) – Dataset of tuples. [0] true dataset, [1] generator input dataset.

  • generator_model (tf.keras.Model) – The generator.

  • discriminator_model (tf.keras.Model) – The discriminator.

  • encoder_model (tf.keras.Model) – The encoder.

  • generator_loss (ashpy.losses.Executor()) – The generator loss.

  • discriminator_loss (ashpy.losses.Executor()) – The discriminator loss.

  • encoder_loss (ashpy.losses.Executor()) – The encoder loss.

  • metrics (list of [ashpy.metrics.metric.Metric]) – All the metrics to be used to evaluate the model.

  • log_eval_mode (ashpy.modes.LogEvalMode) – Models’ mode to use when evaluating and logging.

  • global_step (tf.Variable) – tf.Variable that keeps track of the training steps.

  • ckpt (tf.train.Checkpoint) – checkpoint to use to keep track of models status.

Return type

None

property encoder_loss

Retrieve the encoder loss.

Return type

Optional[Executor]

property encoder_model

Retrieve the encoder model.

Return type

Model

Returns

tf.keras.Model.