CRASpell

The code for our ACL2022 findings paper: CRACSpell: A Contextual Typo Robust Approach with Copy Mechanism to Improve Chinese Spelling Correction

Generate Convert Improve

Install / Use

/learn @liushulinle/CRASpell

About this skill

Quality Score

0/100

README

CRASpell

Source code for the paper "CRASpell: A Contextual Typo Robust Approach to Improve Chinese Spelling Correction" in ACL2022 findings https://aclanthology.org/2022.findings-acl.237.pdf .

1. Requirements

-python 3.7

-tensorflow 1.14

2. Instructions

Step1: Download the pretrained cBERT from https://drive.google.com/file/d/1cqSTpn7r9pnDcvMoM3BbX1X67JsPdZ8_/view?usp=sharing (our previous work), 
and save it in ./datas/init_bert/cbert

Step2: Run the training script: sh start_train.sh
      The best model will be saved when it is finished.

Step3: Run the evaluation script to obtain the results on whole set and multi-typo set, respectively:
      sh start_eval.sh sighan15_test.sh
      sh start_eval.sh sighan15_multierror.txt
      sh start_eval.sh sighan14_test.sh
      sh start_eval.sh sighan14_multierror.txt

Related Skills

node-connect

349.0k

Diagnose OpenClaw node connection and pairing failures for Android, iOS, and macOS companion apps

frontend-design

109.4k

Create distinctive, production-grade frontend interfaces with high design quality. Use this skill when the user asks to build web components, pages, or applications. Generates creative, polished code that avoids generic AI aesthetics.

openai-whisper-api

349.0k

Transcribe audio via OpenAI Audio Transcriptions API (Whisper).

qqbot-media

349.0k

QQBot 富媒体收发能力。使用 <qqmedia> 标签，系统根据文件扩展名自动识别类型（图片/语音/视频/文件）。