pyopmnearwell.ml.scaler_layers module#

Provide MinMax scaler layers for tensorflow.keras.

Warning: Tensorflow 2.17 and Keras 3.0 introduce many pylint errors, hence we disable linting completely. It is possible that the module is not functional at the moment.

class ScalerLayer(*args, **kwargs)[source]#

Bases: Layer

MixIn to provide functionality for the Scaler Layer.

Parameters:

data_min (Tensor)
data_max (Tensor)
feature_range (Sequence[float] | ndarray | Tensor)

min: Tensor#

scalar: Tensor#

build(input_shape)[source]#

Initialize scaling bounds when the layer is built.

not been initialized yet.

Parameters:: input_shape (tuple[int, ...]) -- Shape of input tensors presented to the layer.
Return type:: None

get_weights()[source]#

Return parameters of the scaling.

Returns:: List with three elements in the following order: self.data_min_, self.data_max_, self.feature_range_
Return type:: list[ArrayLike]

set_weights(weights)[source]#

Set parameters of the scaling.

Parameters:

weights (list[ArrayLike]) -- List with three elements in the following order:
data_min
data_max
feature_range

Raises:

ValueError -- If feature_range[0] >= feature_range[1].

Return type:

None

adapt(data)[source]#

Fit scaling bounds independently for each input feature.

each input feature.

Parameters:: data (ArrayLike) -- Values used to fit the scaling extrema.
Return type:: None

_adapt()[source]#

Derive scaling offsets and ranges from the configured extrema.

Return type:: None

property is_adapted#

Return whether the scaling parameters have been initialized.

Returns:: Result produced by the operation.
Return type:: Any

property feature_range#

Return the target interval for each scaled feature.

Returns:: Result produced by the operation.
Return type:: Any

property data_min#

Return the observed minimum for each feature.

Returns:: Result produced by the operation.
Return type:: Any

property data_max#

Return the observed maximum for each feature.

Returns:: Result produced by the operation.
Return type:: Any

class MinMaxScalerLayer(*args, **kwargs)[source]#

Bases: ScalerLayer, Layer

Scales the input according to MinMaxScaling.

See https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.MinMaxScaler.html for an explanation of the transform.

Parameters:

data_min (Tensor)
data_max (Tensor)
feature_range (Sequence[float] | ndarray | Tensor)

name: str#

call(inputs)[source]#

Apply the layer transformation to an input tensor.

Parameters:: inputs (tf.Tensor) -- Tensor values transformed by the layer.
Returns:: Result produced by the operation.
Return type:: tf.Tensor

compute_output_shape(input_shape)[source]#

Return the unchanged output shape of the elementwise transformation.

Parameters:: input_shape (Any) -- Shape of input tensors presented to the layer.
Returns:: Result produced by the operation.
Return type:: Any

get_config()[source]#

Return a serializable Keras layer configuration.

Returns:: Result produced by the operation.
Return type:: Any

classmethod from_config(config)[source]#

Reconstruct a scaler layer from serialized configuration.

Parameters:: config (Any) -- Serialized Keras layer configuration.
Returns:: Result produced by the operation.
Return type:: Any

class MinMaxUnScalerLayer(*args, **kwargs)[source]#

Bases: ScalerLayer, Layer

Unscales the input by applying the inverse transform of MinMaxScalerLayer.

See https://scikit-learn.org/stable/modules/generated/sklearn.preprocessing.MinMaxScaler.html for an explanation of the transformation.

Parameters:

data_min (Tensor)
data_max (Tensor)
feature_range (Sequence[float] | ndarray | Tensor)

_name: str#

call(inputs)[source]#

Apply the layer transformation to an input tensor.

Parameters:: inputs (tf.Tensor) -- Tensor values transformed by the layer.
Returns:: Result produced by the operation.
Return type:: tf.Tensor

compute_output_shape(input_shape)[source]#

Return the unchanged output shape of the elementwise transformation.

Parameters:: input_shape (Any) -- Shape of input tensors presented to the layer.
Returns:: Result produced by the operation.
Return type:: Any

get_config()[source]#

Return a serializable Keras layer configuration.

Returns:: Result produced by the operation.
Return type:: Any

classmethod from_config(config)[source]#

Reconstruct a scaler layer from serialized configuration.

Parameters:: config (Any) -- Serialized Keras layer configuration.
Returns:: Result produced by the operation.
Return type:: Any