tgea's Introduction

TGEA 2.0

Datasets and codes for the paper "TGEA 2.0: A Large-Scale Diagnostically Annotated Dataset with Benchmark Tasks for Text Generation of Pretrained Language Models".

Data License

Attribution-ShareAlike 4.0 International (CC BY-SA 4.0) license. (License URL: https://creativecommons.org/licenses/by-sa/4.0/)

Quick Start

Data Preprocessing

Converting raw data to the format of each task

unzip data.zip
python data/convert_raw_data_to_benchmarks.py 
python data/convert_gec_format.py

Benchmarks

Erroneous Text Detection

sh Diagnosis_tasks/train_b1.sh

MiSEW Extraction

sh Diagnosis_tasks/train_b2.sh

Erroneous Span Location

sh Diagnosis_tasks/train_b3.sh

Error Type Classification

sh Diagnosis_tasks/train_b4.sh

Error Correction

sh Diagnosis_tasks/train_b5.sh

m2scorer is used to evaluate results of error correction.

Generation Pathology Mitigation

sh Generation_Pathology_Mitigation/train_b6.sh
python Generation_Pathology_Mitigation/evaluate.py

tgea's People

Contributors

Stargazers

Watchers

tgea's Issues

Problem with loading dataset using load_dataset

Hi, I have encountered an issue that the dataset you published on huggingface (https://huggingface.co/datasets/jerma66/TGEA2.0) cannot be loaded using the load_dataset tool in python. Relavent code is simplified as below:

from datasets import load_dataset
if __name__ == '__main__':
    dataset = load_dataset('jerma66/TGEA2.0') # ERROR THROWN

Part of the error information is shown below:

...
TypeError: Couldn't cast array of type
struct<错误位置: string, 错误原因: string, 错误大类: string, 错误小类: string, 错误词集位置: string, 错误词集: string, 修正词: string>
to
null

The above exception was the direct cause of the following exception:

...
    raise DatasetGenerationError("An error occurred while generating the dataset") from e
datasets.builder.DatasetGenerationError: An error occurred while generating the dataset

Could you fix this issue?

Recommend Projects