`R/learning_rate_schedules.R`

`learning_rate_schedule_cosine_decay.Rd`

A LearningRateSchedule that uses a cosine decay schedule

```
learning_rate_schedule_cosine_decay(
initial_learning_rate,
decay_steps,
alpha = 0,
...,
name = NULL
)
```

- initial_learning_rate
A scalar

`float32`

or`float64`

Tensor or a R number. The initial learning rate.- decay_steps
A scalar

`int32`

or`int64`

`Tensor`

or an R number. Number of steps to decay over.- alpha
A scalar

`float32`

or`float64`

Tensor or an R number. Minimum learning rate value as a fraction of initial_learning_rate.- ...
For backwards and forwards compatibility

- name
String. Optional name of the operation. Defaults to 'CosineDecay'.

See Loshchilov & Hutter, ICLR2016, SGDR: Stochastic Gradient Descent with Warm Restarts.

When training a model, it is often useful to lower the learning rate as
the training progresses. This schedule applies a cosine decay function
to an optimizer step, given a provided initial learning rate.
It requires a `step`

value to compute the decayed learning rate. You can
just pass a TensorFlow variable that you increment at each training step.

The schedule is a 1-arg callable that produces a decayed learning rate when passed the current optimizer step. This can be useful for changing the learning rate value across different invocations of optimizer functions. It is computed as:

```
decayed_learning_rate <- function(step) {
step <- min(step, decay_steps)
cosine_decay = <- 0.5 * (1 + cos(pi * step / decay_steps))
decayed <- (1 - alpha) * cosine_decay + alpha
initial_learning_rate * decayed
}
```

Example usage:

```
decay_steps <- 1000
lr_decayed_fn <-
learning_rate_schedule_cosine_decay(initial_learning_rate, decay_steps)
```

You can pass this schedule directly into a keras Optimizer
as the `learning_rate`

.