# Model description We pretrained a RoBERTa-based Japanese masked language model on paper abstracts from the academic database CiNii Articles. [A Japanese Masked Language Model for Academic Domain](https://aclanthology.org/2022.sdp-1.16/) # Vocabulary The vocabulary consists of 32000 tokens including subwords induced by the unigram language model of sentencepiece. --- license: apache-2.0
language:ja ---