Normalizers
These are pipeline components that transform the input format or content while preserving their original domain semantics. Think of normalizers as necessary pre-processors that standardize data for downstream steps. Some types of normalizers include:
Type | Description |
---|---|
scalers | adjust values within a target range ie: MinMaxScaler , ImageCropper |
converters | modify values to a common range ie: LanguageTranslator , CurrencyConverter |
formatters | change the structure of the object ie: PDFConverter |
enrichers | add context or details to the objects ie: NameEntityTagger , Q&ALLM |
scikitlab.normalizers.sparsity.DenseTransformer
DenseTransformer()
Converts a sparse matrix into a dense format. This may consume lots of memory.
scikitlab.normalizers.sparsity.SparseTransformer
SparseTransformer()
Converts a dense matrix of data into a compressed sparse row format to preserve memory.