emmi.pipeline.sample_processors.samplewise_normalization

Classes

SamplewiseNormalizationSampleProcessor

Normalizes samplewise to a specified range.

Module Contents

class emmi.pipeline.sample_processors.samplewise_normalization.SamplewiseNormalizationSampleProcessor(item, low, high)

Bases: ksuit.data.pipeline.SampleProcessor

Normalizes samplewise to a specified range.

Parameters:

• item (str) – The item to normalize.

• low (collections.abc.Sequence[float] | None) – The low value of the range.

• high (collections.abc.Sequence[float]) – The high value of the range.

item

high_tensor

abstractmethod inverse(key, value)

Denormalization is not implemented.

Parameters:

• key (str)

• value (torch.Tensor)

Return type:

tuple[str, torch.Tensor]