roberta-base-japanese-char-luw-upos

Model Description

This is a RoBERTa model pre-trained on 青空文庫 texts for POS-tagging and dependency-parsing, derived from roberta-base-japanese-aozora-char. Every long-unit-word is tagged by UPOS (Universal Part-Of-Speech) and FEATS.

How to Use

from transformers import AutoTokenizer,AutoModelForTokenClassification,TokenClassificationPipeline
tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/roberta-base-japanese-char-luw-upos")
model=AutoModelForTokenClassification.from_pretrained("KoichiYasuoka/roberta-base-japanese-char-luw-upos")
pipeline=TokenClassificationPipeline(tokenizer=tokenizer,model=model,aggregation_strategy="simple")
nlp=lambda x:[(x[t["start"]:t["end"]],t["entity_group"]) for t in pipeline(x)]
print(nlp("国境の長いトンネルを抜けると雪国であった。"))

or

import esupar
nlp=esupar.load("KoichiYasuoka/roberta-base-japanese-char-luw-upos")
print(nlp("国境の長いトンネルを抜けると雪国であった。"))

Reference

安岡孝一: Transformersと国語研長単位による日本語係り受け解析モデルの製作, 情報処理学会研究報告, Vol.2022-CH-128, No.7 (2022年2月), pp.1-8.

See Also

esupar: Tokenizer POS-tagger and Dependency-parser with BERT/RoBERTa/DeBERTa models