K

Kcbert Base

Developed by beomi
KcBERT is a BERT model specifically optimized for Korean comment scenarios, trained on Naver news comment data, excelling in handling informal expressions and internet slang.
Downloads 82.60k
Release Time : 3/2/2022

Model Overview

KcBERT is a pre-trained language model optimized for Korean user comment scenarios (such as movie reviews and social comments). It is trained on a large collection of news comment data and can effectively handle informal expressions, internet neologisms, and spelling errors in Korean.

Model Features

Korean Comment Optimization
Specifically trained for Korean user comment scenarios, effectively handling informal expressions, internet neologisms, and spelling errors
Complete Training Process
Tokenizer and model trained from scratch, not fine-tuned from existing models
Excellent Multi-task Performance
Outperforms general Korean BERT models on Korean NLP tasks such as NSMC movie review classification

Model Capabilities

Korean Text Understanding
Sentiment Analysis
Named Entity Recognition
Semantic Similarity Calculation
Masked Language Modeling

Use Cases

Sentiment Analysis
Movie Review Classification
Classifying Naver movie reviews as positive/negative
Base model 89.62% accuracy, Large model 90.68% accuracy
Text Understanding
News Comment Analysis
Understanding user comments containing internet slang and spelling errors
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase