Skip to content
This repository has been archived by the owner on Mar 18, 2024. It is now read-only.

Added model file - CallmeMehdi #66

Merged
merged 12 commits into from
Aug 9, 2021
13 changes: 13 additions & 0 deletions assets/docs/callmemehdi/callmemehdi.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,13 @@
# Publisher callmemehdi
Mehdi Mabrouki.

[![Icon URL]](https://avatars.githubusercontent.com/u/47258917?v=4)

## Details
Mehdi Mabrouki ([@CallmeMehdi](https://github.com/CallmeMehdi)).
A Tunisian Software Engineer and Artificial Intelligence enthusiast, interested in Machine and Deep Learning.
Google Summer of Code 2021 participant.

**GitHub:** [@CallmeMehdi](https://github.com/CallmeMehdi)\
**LinkedIn:** [Mehdi Mabrouki](https://www.linkedin.com/in/mehdimabrouki/)\
**Personal Website:** [portfolio](https://portfolio.callmemehdi.vercel.app/)
44 changes: 44 additions & 0 deletions assets/docs/callmemehdi/models/AraBERT/1.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,44 @@
# Module callmemehdi/arabert/1
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please update the 'arabert' casing here or in the folder name so that they match


AraBERT is an Arabic pretrained lanaguage model based on Google's BERT architechture.

<!-- asset-path: https://gsoctfarabert.web.app/arabert.tar.gz -->
<!-- task: text-embedding -->
<!-- fine-tunable: true -->
<!-- format: saved_model_2 -->
<!-- license: custom -->

## Overview

These results are obtained from pre-training BERT specifically for the Arabic language in the pursuit of achieving the same success that BERT did for the English language.The performance of AraBERT is compared to multilingual BERT from Google and other state-of-the-art approaches
MorganR marked this conversation as resolved.
Show resolved Hide resolved

## Performance

Performance of AraBERT on Arabic downstreamtasks compared to mBERT
Results:

Task | metric | mBERT | AraBERT
------|--------|------ | -------
SA (HARD) | Acc. | 95.7 | 96.1
SA (ASTD) | Acc. | 80.1 | 96.5
SA (ArsenTD-Lev) | Acc. | 51.0 | 59.4
SA (AJGT) | Acc. | 83.6 | 93.8
SA (LABR) | Acc. | 83.0 | 86.7
NER (ANERcorp) | macro-F1 | 78.4 | 81.9
QA (ARCD) | macro-F1 | 61.3 | 62.7

MorganR marked this conversation as resolved.
Show resolved Hide resolved

## Reference

```
@inproceedings{antoun2020arabert,
title={AraBERT: Transformer-based Model for Arabic Language Understanding},
author={Antoun, Wissam and Baly, Fady and Hajj, Hazem},
booktitle={LREC 2020 Workshop Language Resources and Evaluation Conference 11--16 May 2020},
pages={9}
}
```

## License

[SOFTWARE LICENSE AGREEMENT - AraBERT](https://github.com/snakers4/silero-models/blob/master/LICENSE)
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes, absolutely, I think I didn't save when I changed it. Changing it right now.