Sikuroberta
A pre-trained language model for classical Chinese text processing, trained on the full corpus of 'Siku Quanshu'
Downloads 326
Release Time : 3/2/2022
Model Overview
Siku BERT and Siku RoBERTa are pre-trained language models based on the BERT architecture, specifically designed for intelligent processing tasks of classical Chinese, trained with high-quality 'Siku Quanshu' corpus
Model Features
Specialized Classical Chinese Processing
Pre-trained model optimized specifically for the characteristics of classical Chinese
High-Quality Corpus Training
Trained with the verified full text of 'Siku Quanshu' as training data
BERT Architecture Optimization
Optimized based on the BERT deep language model architecture
Model Capabilities
Classical text understanding
Classical text information extraction
Classical text mining
Use Cases
Digital Humanities Research
Ancient Text Analysis
Semantic analysis and information extraction of ancient texts
Classical Text Automatic Processing
Supports various classical Chinese natural language processing tasks
Featured Recommended AI Models
Š 2025AIbase