Abstract base class for parameter regularization.
Provides the interface for implementing regularization terms that can be
added to the training loss. Subclasses must implement loss, sample_init,
and reset_value methods.
- Parameters:
fit_hyper (bool) – Whether to optimize the hyperparameters of the regularization
as trainable parameters. Default is False.
-
fit_hyper
Whether hyperparameters are trainable.
- Type:
bool
Notes
Regularization can be used with the Param class to add regularization
terms to the training loss.
-
loss(value)[source]
Calculate regularization loss.
- Parameters:
value (Array | ndarray | bool | number | bool | int | float | complex | Quantity) – Parameter values to compute regularization for.
- Returns:
Scalar regularization loss.
- Return type:
Array | ndarray | bool | number | bool | int | float | complex | Quantity
-
reset_value()[source]
Return the reset value (e.g., prior mean).
- Returns:
Value to reset the parameter to.
- Return type:
Array | ndarray | bool | number | bool | int | float | complex | Quantity
-
sample_init(shape)[source]
Sample initial value from the regularization’s implied prior distribution.
- Parameters:
shape (int | Sequence[int] | integer | Sequence[integer]) – Shape of the parameter to initialize.
- Returns:
Sampled initial value.
- Return type:
Array | ndarray | bool | number | bool | int | float | complex | Quantity
