TextSETTR Label Free Text Style Extraction and Tunable Targeted Restyling

The technique removes the need for labels entirely, relying instead on the implicit connection in style between adjacent sentences in unlabeled text . We show that a strong pretrained text-to-text model can be adapted to extract a style vector from arbitrary text and use this vector to condition the decoder to perform style transfer . The resulting learned style vector space encodes many facets of textual style, we recast transfers as “targeted restyling” vector operations that adjust specific attributes of the input text while preserving others . When trained over Amazon reviews data, our resulting TextSETTR model is competitive on sentiment transfer, even when given only four exemplars of each class . Furthermore, we demonstrate that a single model trained on Common Crawl data is capable of transferring along multiple dimensions including dialect, emotiveness, form

Links: PDF - Abstract

Code :


Keywords : style - text - vector - model - transfer -

Leave a Reply

Your email address will not be published. Required fields are marked *