AIbase
EN
Home
Roberta Base Japanese Aozora
R

Roberta Base Japanese Aozora

Developed by KoichiYasuoka
Japanese RoBERTa model pre-trained on Aozora Bunko texts, supporting masked language modeling tasks
Large Language Model
Transformers
 Japanese#Japanese Text Processing#Aozora Bunko Pre-training#Dependency Parsing
Downloads 17
Release Time : 3/2/2022

Model Overview

This is a RoBERTa model pre-trained on Japanese Aozora Bunko texts using Japanese-LUW-Tokenizer, suitable for Japanese text processing tasks such as part-of-speech tagging and dependency parsing.

Model Features

Aozora Bunko Pre-training
Pre-trained on classic literary works from Japan's Aozora Bunko, suitable for processing formal Japanese texts
Dedicated Japanese Tokenizer
Uses Japanese-LUW-Tokenizer for word segmentation, optimized for Japanese language characteristics
Downstream Task Adaptability
Can be fine-tuned for various Japanese NLP tasks such as part-of-speech tagging and dependency parsing

Model Capabilities

Japanese text understanding
Masked language modeling
Part-of-speech tagging
Dependency parsing

Use Cases

Natural Language Processing
Japanese Text Completion
Predicts masked words in sentences
Example: 'Upon arriving in Japan, visit [MASK].' can predict appropriate location nouns
Grammar Analysis
Used for Japanese part-of-speech tagging and dependency relation analysis
There are fine-tuned versions of this model for part-of-speech tagging and dependency parsing
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
English简体中文繁體中文にほんご
© 2025AIbase