Schedule-free AdamW optimizer

Schedule-free learning avoids a separate learning-rate schedule by combining interpolation and averaging. It removes the need to specify the stopping time in advance and typically matches or outperforms cosine and linear decay schedules. The optimizer maintains a momentum sequence to which gradient updates are applied and an averaged sequence used for evaluation. During training, model parameters interpolate between the two sequences.

Usage

optimizer_schedule_free_adam_w(
  learning_rate = 0.0025,
  beta_1 = 0.9,
  beta_2 = 0.999,
  epsilon = 1e-08,
  warmup_steps = 0L,
  weight_decay = NULL,
  clipnorm = NULL,
  clipvalue = NULL,
  global_clipnorm = NULL,
  use_ema = FALSE,
  ema_momentum = 0.99,
  ema_overwrite_frequency = NULL,
  loss_scale_factor = NULL,
  gradient_accumulation_steps = NULL,
  name = NULL,
  ...
)

Arguments

learning_rate: A number, a LearningRateSchedule() instance, or a callable that takes no arguments and returns the value to use. Defaults to 0.0025.
beta_1: Number, constant tensor, or callable returning the exponential decay rate for the first-moment estimates. It also controls interpolation between the momentum and averaged sequences. Defaults to 0.9.
beta_2: A float value or a constant float tensor, or a callable that takes no arguments and returns the actual value to use. The exponential decay rate for the 2nd moment estimates. Defaults to 0.999.
epsilon: Small constant for numerical stability. Defaults to 1e-8.
warmup_steps: Number of warmup steps. During warmup, the learning rate increases linearly from zero to learning_rate. Defaults to 0.
weight_decay: Float. If set, weight decay is applied.
clipnorm: Float. If set, the gradient of each weight is individually clipped so that its norm is no higher than this value.
clipvalue: Float. If set, the gradient of each weight is clipped to be no higher than this value.
global_clipnorm: Float. If set, the gradient of all weights is clipped so that their global norm is no higher than this value.
use_ema: Boolean, defaults to FALSE. If TRUE, exponential moving average (EMA) is applied. EMA consists of computing an exponential moving average of the weights of the model (as the weight values change after each training batch), and periodically overwriting the weights with their moving average.
ema_momentum: Float, defaults to 0.99. Only used if use_ema=TRUE. This is the momentum to use when computing the EMA of the model's weights: new_average = ema_momentum * old_average + (1 - ema_momentum) * current_variable_value.
ema_overwrite_frequency: Int or NULL, defaults to NULL. Only used if use_ema=TRUE. Every ema_overwrite_frequency steps of iterations, we overwrite the model variable by its moving average. If NULL, the optimizer does not overwrite model variables in the middle of training, and you need to explicitly overwrite the variables at the end of training by calling optimizer$finalize_variable_values() (which updates the model variables in-place). When using the built-in fit() training loop, this happens automatically after the last epoch, and you don't need to do anything.
loss_scale_factor: Float or NULL. If a float, the scale factor will be multiplied the loss before computing gradients, and the inverse of the scale factor will be multiplied by the gradients before updating variables. Useful for preventing underflow during mixed precision training. Alternately, optimizer_loss_scale() will automatically set a loss scale factor.
gradient_accumulation_steps: Int or NULL. If an int, model and optimizer variables will not be updated at every step; instead they will be updated every gradient_accumulation_steps steps, using the average value of the gradients since the last update. This is known as "gradient accumulation". This can be useful when your batch size is very small, in order to reduce gradient noise at each update step. EMA frequency will look at "accumulated" iterations value (optimizer steps // gradient_accumulation_steps). Learning rate schedules will look at "real" iterations value (optimizer steps).
name: String. The name to use for momentum accumulator weights created by the optimizer.
...: For forward/backward compatibility.

Value

An Optimizer instance.

Examples

optimizer <- optimizer_schedule_free_adam_w(learning_rate = 0.0025)
model |> compile(optimizer = optimizer, loss = "mse")
model |> fit(x_train, y_train)

Usage

Arguments

Value

Examples

References

See also