---
language:
- "ja"
tags:
- "japanese"
- "masked-lm"
- "modernbert"
datasets:
- "globis-university/aozorabunko-clean"
- "wikimedia/wikipedia"
license: "apache-2.0"
pipeline_tag: "fill-mask"
mask_token: "[MASK]"
widget:
- text: "日本に着いたら[MASK]を訪ねなさい。"
---

# modernbert-base-japanese-wikipedia

## Model Description

This is a ModernBERT model pre-trained on Japanese Wikipedia and 青空文庫 texts. NVIDIA A100-SXM4-40GB×8 took 56 hours 49 minutes for training. You can fine-tune `modernbert-base-japanese-wikipedia` for downstream tasks, such as [POS-tagging](https://huggingface.co/KoichiYasuoka/modernbert-base-japanese-wikipedia-upos), [dependency-parsing](https://huggingface.co/KoichiYasuoka/modernbert-base-japanese-wikipedia-ud-square), and so on.

## How to Use

```py
from transformers import AutoTokenizer,AutoModelForMaskedLM
tokenizer=AutoTokenizer.from_pretrained("KoichiYasuoka/modernbert-base-japanese-wikipedia")
model=AutoModelForMaskedLM.from_pretrained("KoichiYasuoka/modernbert-base-japanese-wikipedia",trust_remote_code=True)
```

### Reference

安岡孝一: [青空文庫ModernBERTモデルによる国語研長単位係り受け解析](http://hdl.handle.net/2433/291645), 情報処理学会研究報告, Vol.2025-CH-137『人文科学とコンピュータ』, No.10 (2025年2月8日), pp.1-7.