R

Roformer V2 Chinese Char Large

Developed by junnyu
RoFormerV2 is an enhanced Transformer model based on rotary position embedding, developed by Zhuiyi Technology, supporting Chinese text processing tasks.
Downloads 84
Release Time : 3/21/2022

Model Overview

RoFormerV2 is an improved Transformer model utilizing rotary position embedding technology, suitable for various Chinese natural language processing tasks such as text classification, question answering, and language understanding.

Model Features

Rotary Position Embedding
Utilizes rotary position embedding technology, enhancing the model's ability to capture positional information.
Multi-task Learning
Supports multi-task learning, improving performance across various tasks.
Improved Classification Head
Added 2 dropout layers, 1 fully connected layer, and ReLU activation function to the original code, enhancing classification performance.

Model Capabilities

Text classification
Question answering
Language understanding
Text generation

Use Cases

Natural Language Processing
Text Classification
Performs excellently in CLUE development set classification tasks, such as datasets from iFLYTEK and Tencent News.
Outperforms models like BERT and RoBERTa on multiple datasets.
Question Answering System
Suitable for building Chinese question answering systems, capable of understanding and answering user questions.
Featured Recommended AI Models
AIbase
Empowering the Future, Your AI Solution Knowledge Base
© 2025AIbase