Methods › Natural Language Processing › Tokenizers › SentencePiece

SentencePiece

908 papers tagged archive 2025-07-28

archive 2025-07-28 Description, source and code snippet are the archive's method entry.

SentencePiece is a subword tokenizer and detokenizer for natural language processing. It performs subword segmentation, supporting the byte-pair-encoding (BPE) algorithm and unigram language model, and then converts this text into an id sequence guarantee perfect reproducibility of the normalization and subword segmentation.

Source: SentencePiece: A simple and language independent subword...

Papers archive 2025-07-28

30 shown of 908, newest first. Repository counts are the archive's code-links table. A Syntology line states what Syntology ran from that paper's harvested code; it is per sample and not a correctness claim.

Tasks archive 2025-07-28

20 shown of 560 tasks the archive attaches to papers tagged with this method, by distinct papers. A task without a page in the catalog is plain text.

TaskPapers
Language Modelling154
Language Modeling123
Question Answering104
Sentence82
Decoder78
Text Generation73
Translation61
Machine Translation55
Transfer Learning54
Retrieval47
Sentiment Analysis41
Natural Language Understanding37
Data Augmentation29
Text Classification29
Natural Language Inference26
Reading Comprehension24
text-classification24
Abstractive Text Summarization23
Semantic Parsing23
Articles21

Usage over time archive 2025-07-28

Papers per year tagged with SentencePiece: 2018 to 2025, peak 219 219 0 2018: 1 paper 2018 2019: 27 papers 2019 2020: 97 papers 2020 2021: 154 papers 2021 2022: 185 papers 2022 2023: 219 papers 2023 2024: 174 papers 2024 2025: 51 papers 2025
Papers per year the archive tags with this method, by the paper's archive date (908 dated). Bars are counts, not a trend claim.

Components: the archive holds no method-to-method composition, so PwC's Components table cannot be rebuilt; the Papers list carries no Results column for the same reason (the archive does not join its leaderboard rows to method tags).

Categories archive 2025-07-28

Tokenizers

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections