org.platanios.tensorflow.api.ops.training.optimizers.schedules
Cosine decay cycle in terms of number of steps.
Minimum decayed learning rate value as a fraction of the original learning rate value.
Step after which to start decaying the learning rate.
Composes the provided other
schedule with this schedule and returns the resulting schedule.
Composes the provided other
schedule with this schedule and returns the resulting schedule.
Minimum decayed learning rate value as a fraction of the original learning rate value.
Applies the decay method to value
, the current iteration in the optimization loop is step
and returns the
result.
Applies the decay method to value
, the current iteration in the optimization loop is step
and returns the
result.
Value to decay.
Option containing current iteration in the optimization loop, if one has been provided.
Decayed value.
IllegalArgumentException
If the decay method requires a value for step
but the provided option is empty.
Composes this schedule with the provided, other
schedule and returns the resulting schedule.
Composes this schedule with the provided, other
schedule and returns the resulting schedule.
Cosine decay cycle in terms of number of steps.
Step after which to start decaying the learning rate.
Cosine decay method.
This method applies a cosine decay function to a provided initial learning rate (i.e.,
value
). It requires a step value to be provided in it's application function, in order to compute the decayed learning rate. You may simply pass a TensorFlow variable that you increment at each training step.The decayed value is computed as follows: