PROTECT: Parameter-Efficient Tuning for Few-Shot Robust Chinese Text Correction

---

This repository contains code, model, datasets for PROTECT

cd scripts
sh pretrain.sh

Finetune

pretrained model and datasets

To facilitate replication, we provide pre-trained PROTECT model and datasets.

pre-trained PROTECT model available at here

datasets available at here

cd scripts
sh train.sh

Citation

Xuan Feng, Tianlong Gu, Liang Chang, Xiaoli Liu, PROTECT: Parameter-Efficient Tuning for Few-Shot Robust Chinese Text Correction[J], in IEEE/ACM Transactions on Audio, Speech, and Language Processing (TASLP 2024), doi: 10.1109/TASLP.2024.3414333.