GitHub - lingochamp/Multi-Scale-BERT-AES: Demo for the paper "On the Use of BERT for Automated Essay Scoring: Joint Learning of Multi-Scale Essay Representation"

lingochamp / Multi-Scale-BERT-AES Public

Notifications You must be signed in to change notification settings
Fork 13
Star 56

Demo for the paper "On the Use of BERT for Automated Essay Scoring: Joint Learning of Multi-Scale Essay Representation"

56 stars 13 forks Branches Tags Activity

Notifications

Name		Name	Last commit message	Last commit date
Latest commit History 2 Commits
data		data
Readme		Readme
asap.ini		asap.ini
data.py		data.py
decode.sh		decode.sh
document_bert_architectures.py		document_bert_architectures.py
encoder.py		encoder.py
evaluate.py		evaluate.py
model_architechure_bert_multi_scale_multi_loss.py		model_architechure_bert_multi_scale_multi_loss.py
pred.txt		pred.txt
predict_multi_scale_multi_loss.py		predict_multi_scale_multi_loss.py
requirements.txt		requirements.txt

Repository files navigation

This is a demo for the paper(NAACL 2022 long paper accepted): "On the Use of BERT for Automated Essay Scoring: Joint Learning of Multi-Scale Essay Representation"

Steps to run the decoding process:
step 1: Download the project to the computer.
step 2: Download the model(for ASAP's prompt 8) to the computer with the following link:
        Link: https://pan.baidu.com/s/1_m_-DQlX-dLh1XdhOMzj1A?pwd=tmmb
step 3: Update the config file "asap.ini" and set the parameters as following:
        data_dir: your local directory to store the essays of prompt 8 (file format: id \t text \t score);
        model_directory: your local directory to store the model which is obtained in step 2;
        result_file: the path of file to store prediction result.
step 4: Run the following script to get scoring result:
        sh decode.sh