TensorFlow 2.0 Beta is available Learn more

Module: tf.keras.experimental

Aliases:

  • Module tf.compat.v1.keras.experimental
  • Module tf.keras.experimental

Defined in python/keras/api/_v1/keras/experimental/__init__.py.

Classes

class CosineDecay: A LearningRateSchedule that uses a cosine decay schedule.

class CosineDecayRestarts: A LearningRateSchedule that uses a cosine decay schedule with restarts.

class LinearCosineDecay: A LearningRateSchedule that uses a linear cosine decay schedule.

class NoisyLinearCosineDecay: A LearningRateSchedule that uses a noisy linear cosine decay schedule.

class PeepholeLSTMCell: Equivalent to LSTMCell class but adds peephole connections.

class SequenceFeatures: A layer for sequence input.

Functions

export_saved_model(...): Exports a tf.keras.Model as a Tensorflow SavedModel.

load_from_saved_model(...): Loads a keras Model from a SavedModel created by export_saved_model().

terminate_keras_multiprocessing_pools(...): Destroy Keras' multiprocessing pools to prevent deadlocks.