SentencePiece: A simple and language independent subword tokenizer and detokenizer for Neural Text Processing
This provides a purely end-to-end and language-independent tokenization solution for neural-based text processing, such as machine translation, though it is incremental relative to existing subword tools.
The paper tackles the problem of language-independent subword tokenization for neural text processing by introducing SentencePiece, a tool that trains subword models directly from raw sentences, achieving comparable accuracy to existing methods in English-Japanese machine translation.
This paper describes SentencePiece, a language-independent subword tokenizer and detokenizer designed for Neural-based text processing, including Neural Machine Translation. It provides open-source C++ and Python implementations for subword units. While existing subword segmentation tools assume that the input is pre-tokenized into word sequences, SentencePiece can train subword models directly from raw sentences, which allows us to make a purely end-to-end and language independent system. We perform a validation experiment of NMT on English-Japanese machine translation, and find that it is possible to achieve comparable accuracy to direct subword training from raw sentences. We also compare the performance of subword training and segmentation with various configurations. SentencePiece is available under the Apache 2 license at https://github.com/google/sentencepiece.