Skip to content

Text Detoxification Task is a process of transforming the text with toxic style into the text with the same meaning but with neutral style.

Notifications You must be signed in to change notification settings

karinaDen/text-detoxification

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

14 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Text detoxification Assignment

Innipolis [F23] PML&DL course


🎓 Student

Name: Karina Denisova
Email: [email protected]
Group numer: BS21-DS-01

🗒 Project description

Project aims to address the challenge of transforming text with toxic style into text with a neutral style while preserving the same underlying meaning.

🤾‍♀️ Run the project

  1. Clone the repository
  2. Install requirements
pip install -r requirements.txt
  1. For dataset creation run:
python src/data/make_dataset.py
  1. For baseline model run:
python src/models/baseline/predict_baseline.py 

4.1) For baseline evaluation run:

python src/metrics/metrics.py --inputs data/interim/translation.txt --preds data/interim/baseline.txt
  1. For training T5-small model run:
python src/models/T5/T5_model_train.py

5.1) For T5-small model evaluation run:

python src/metrics/metrics.py --inputs data/interim/test_translation.txt --preds data/interim/result.txt

5.2) To download T5-small model weights run:

python src/data/load_weights.py

Also you can run all modelt and thesting with examples from the notebooks folder.

About

Text Detoxification Task is a process of transforming the text with toxic style into the text with the same meaning but with neutral style.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors