Module: tff.learning.optimizers

Libraries for optimization algorithms.

Classes

class Optimizer: Represents an optimizer for use in TensorFlow Federated.

Functions

build_adafactor(...): Builds an Adafactor optimizer.

build_adagrad(...): Returns a tff.learning.optimizers.Optimizer for Adagrad.

build_adam(...): Returns a tff.learning.optimizers.Optimizer for Adam.

build_adamw(...): Returns a tff.learning.optimizers.Optimizer for AdamW.

build_rmsprop(...): Returns a tff.learning.optimizers.Optimizer for RMSprop.

build_sgdm(...): Returns a tff.learning.optimizers.Optimizer for momentum SGD.

build_yogi(...): Returns a tff.learning.optimizers.Optimizer for Yogi.

check_weights_gradients_match(...): Checks that weights and non-none gradients match.

handle_indexed_slices_gradients(...): Converts any tf.IndexedSlices to tensors.

schedule_learning_rate(...): Returns an optimizer with scheduled learning rate.

LEARNING_RATE_KEY 'learning_rate'