An unsupervised text tokenizer and detokenizer mainly for Neural Network-based text generation systems.
conda install rocketce::sentencepiececonda install rocketce/label/rocketce-1.10.0::sentencepiececonda install rocketce/label/rocketce-1.11.2::sentencepiececonda install rocketce/label/rocketce-1.11.5::sentencepiececonda install rocketce/label/rocketce-1.11.6::sentencepiececonda install rocketce/label/rocketce-1.8.0::sentencepiececonda install rocketce/label/rocketce-1.8.1::sentencepiececonda install rocketce/label/rocketce-1.9.1::sentencepiece
SentencePiece is an unsupervised text tokenizer and detokenizer mainly for Neural Network-based text generation systems where the vocabulary size is predetermined prior to the neural model training. SentencePiece implements subword units (e.g., byte-pair-encoding (BPE) [Sennrich et al.]) and unigram language model [Kudo.]) with the extension of direct training from raw sentences. SentencePiece allows us to make a purely end-to-end system that does not depend on language-specific pre/postprocessing.