https://github.com/michaelgira23/debiasing-lms
Code for the publication Debiasing Pre-Trained Language Models via Efficient Fine-Tuning
https://github.com/michaelgira23/debiasing-lms
Last synced: about 1 month ago
JSON representation
Code for the publication Debiasing Pre-Trained Language Models via Efficient Fine-Tuning
- Host: GitHub
- URL: https://github.com/michaelgira23/debiasing-lms
- Owner: michaelgira23
- Created: 2022-04-09T23:13:20.000Z (about 3 years ago)
- Default Branch: main
- Last Pushed: 2022-05-21T19:26:18.000Z (almost 3 years ago)
- Last Synced: 2025-04-15T00:43:02.397Z (about 1 month ago)
- Language: Python
- Homepage: https://huggingface.co/spaces/michaelgira23/debiasing-lms
- Size: 7.41 MB
- Stars: 5
- Watchers: 2
- Forks: 2
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# ⚖️ Debiasing Language Models
> Official code for _Debiasing Pre-Trained Language Models via Efficient Fine-Tuning_ published in the [Second Workshop on Language Technology for Equality, Diversity, Inclusion](https://sites.google.com/view/lt-edi-2022) at ACL 2022.
[View Demo](https://huggingface.co/spaces/michaelgira23/debiasing-lms) | [View Presentation](https://youtu.be/ErQ2NxIQ9Vg)
**Currently placeholder. Code will be polished and published soon!** In the meantime, you can [take a look at the old code.](https://github.com/michaelgira23/debiasing-lms/tree/old)
# Dataset
Our fine-tuning dataset consists of the [WinoBias](https://github.com/uclanlp/corefBias) and [CrowS-Pairs](https://github.com/nyu-mll/crows-pairs) datasets. After cloning the Git submodules for the respective datasets, run:
```bash
python dataset/prepare.py
```
`prepare.py` combines the datasets from each repository and splits them into a training (80%), cross-validation (10%), and testing sets (10%).