Erya Open-Source Classical Chinese Translation Model - Free Deployment to Assist in Precise Conversion from Classical Chinese to Modern Chinese

Home

Erya

Developed by RUCAIBox

Erya is a pre-trained model specifically designed for translating classical Chinese into modern Chinese.

Machine Translation

Transformers

ChineseOpen Source License:Apache-2.0 #Ancient to Modern Translation #Disyllabic Alignment #Encoder-Decoder

Downloads 17

Release Time : 7/21/2023

Model Overview

Erya adopts an encoder-decoder architecture, combining DMLM and DAS technologies, focusing on the translation task from classical to modern Chinese.

Model Features

Dual Masked Language Model

Utilizes DMLM technology to enhance the model's understanding of classical Chinese.

Disyllabic Alignment Substitution

Employs DAS technology to improve translation accuracy and fluency.

Specialized Dataset

Trained on datasets containing classical and modern Chinese texts to optimize translation performance.

Model Capabilities

Classical Chinese Translation

Modern Chinese Generation

Use Cases

Academic Research

Ancient Text Translation

Translates classical Chinese literature into modern Chinese for easier comprehension by contemporary readers.

Enhances the readability and dissemination efficiency of ancient texts.

Education

Classical Chinese Teaching

Assists teachers and students in understanding classical Chinese texts.

Improves teaching effectiveness and learning efficiency.

🚀 Erya - Ancient Chinese to Modern Chinese Translation Model

Erya is a powerful pretrained model dedicated to translating Ancient Chinese into Modern Chinese, offering high - quality translation solutions.

🚀 Quick Start

If you want to start using Erya for Ancient Chinese to Modern Chinese translation, you can follow the steps below. First, make sure you have installed the necessary libraries. Then, you can use the following code to load the model and perform translation.

✨ Features

Specialized Design: Specifically designed for the task of translating Ancient Chinese into Modern Chinese.
Advanced Architecture: Utilizes an Encoder - Decoder architecture.
Unique Training Techniques: Trained using a combination of DMLM (Dual Masked Language Model) and DAS (Disyllabic Aligned Substitution) techniques on datasets containing both Ancient and Modern Chinese texts.

📦 Installation

Installation steps are not provided in the original document. If you want to use the model, you need to ensure that the transformers library is installed. You can install it using the following command:

pip install transformers

💻 Usage Examples

Basic Usage

>>> from transformers import BertTokenizer, CPTForConditionalGeneration

>>> tokenizer = BertTokenizer.from_pretrained("RUCAIBox/Erya")
>>> model = CPTForConditionalGeneration.from_pretrained("RUCAIBox/Erya")

>>> input_ids = tokenizer("安世字子孺，少以父任为郎。", return_tensors='pt')
>>> input_ids.pop("token_type_ids")

>>> pred_ids = model.generate(max_new_tokens=256, **input_ids)
>>> print(tokenizer.batch_decode(pred_ids, skip_special_tokens=True))
    ['安 世 字 子 孺 ， 年 轻 时 因 父 任 郎 官 。']

📚 Documentation

The detailed information of our work can be found here: RUCAIBox/Erya (github.com). More information about the Erya dataset can be found here: [RUCAIBox/Erya - dataset · Datasets at Hugging Face](https://huggingface.co/datasets/RUCAIBox/Erya - dataset). You can use this dataset to further tune the Erya model for better translation performance.

📄 License

This project is licensed under the Apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご