Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,90 @@
|
|
1 |
---
|
2 |
license: cc-by-4.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: cc-by-4.0
|
3 |
+
language:
|
4 |
+
- tn
|
5 |
+
datasets:
|
6 |
+
- dsfsi/PuoData
|
7 |
+
metrics:
|
8 |
+
- f1
|
9 |
+
library_name: transformers
|
10 |
+
pipeline_tag: text-classification
|
11 |
+
tags:
|
12 |
+
- iptc
|
13 |
---
|
14 |
+
|
15 |
+
|
16 |
+
# PuoBERTa-News: A Setswana Langauge Model Finetuned for News Categorisation
|
17 |
+
|
18 |
+
A Roberta-based language model finetuned for News Categorisation.
|
19 |
+
|
20 |
+
Based on [https://huggingface.co/dsfsi/PuoBERTa](https://huggingface.co/dsfsi/PuoBERTa)
|
21 |
+
|
22 |
+
## Model Details
|
23 |
+
|
24 |
+
### Model Description
|
25 |
+
|
26 |
+
This is a News Categorisation model for Setswana.
|
27 |
+
|
28 |
+
- **Developed by:** Vukosi Marivate ([@vukosi](https://huggingface.co/@vukosi)), Moseli Mots'Oehli ([@MoseliMotsoehli](https://huggingface.co/@MoseliMotsoehli)) , Valencia Wagner, Richard Lastrucci and Isheanesu Dzingirai
|
29 |
+
- **Model type:** RoBERTa Model
|
30 |
+
- **Language(s) (NLP):** Setswana
|
31 |
+
- **License:** CC BY 4.0
|
32 |
+
|
33 |
+
### News Categories
|
34 |
+
|
35 |
+
0: arts_culture_entertainment_and_media
|
36 |
+
1: crime_law_and_justice
|
37 |
+
2: disaster_accident_and_emergency_incident
|
38 |
+
3: economy_business_and_finance
|
39 |
+
4: education
|
40 |
+
5: environment
|
41 |
+
6: health
|
42 |
+
7: politics
|
43 |
+
8: religion_and_belief
|
44 |
+
9: society
|
45 |
+
|
46 |
+
### Model Performance
|
47 |
+
|
48 |
+
Performance of models on Daily News Dikgang dataset
|
49 |
+
|
50 |
+
| **Model** | **5-fold Cross Validation F1** | **Test F1** |
|
51 |
+
|-----------------------------|--------------------------------------|-------------------|
|
52 |
+
| Logistic Regression + TFIDF | 60.1 | 56.2 |
|
53 |
+
| NCHLT TSN RoBERTa | 64.7 | 60.3 |
|
54 |
+
| PuoBERTa | 63.8 | 62.9 |
|
55 |
+
| PuoBERTaJW300 | extbf{66.2} | **65.4**
|
56 |
+
|
57 |
+
### Usage
|
58 |
+
|
59 |
+
Use this model for Part of Speech Tagging for Setswana.
|
60 |
+
```python
|
61 |
+
|
62 |
+
```
|
63 |
+
|
64 |
+
## Citation Information
|
65 |
+
|
66 |
+
Bibtex Reference
|
67 |
+
|
68 |
+
```
|
69 |
+
@article{marivatePuoBERTa2023,
|
70 |
+
title={PuoBERTa: Training and evaluation of a curated language model for Setswana},
|
71 |
+
author={Vukosi Marivate and Moseli Mots'Oehli and Valencia Wagner and Richard Lastrucci and Isheanesu Dzingirai},
|
72 |
+
journal={ArXiv},
|
73 |
+
}
|
74 |
+
```
|
75 |
+
|
76 |
+
## Contributing
|
77 |
+
|
78 |
+
Your contributions are welcome! Feel free to improve the model.
|
79 |
+
|
80 |
+
## Model Card Authors
|
81 |
+
|
82 |
+
Vukosi Marivate
|
83 |
+
|
84 |
+
## Model Card Contact
|
85 |
+
|
86 |
+
For more details, reach out or check our [website](https://dsfsi.github.io/).
|
87 |
+
|
88 |
+
Email: vukosi.marivate@cs.up.ac.za
|
89 |
+
|
90 |
+
**Enjoy exploring Setswana through AI!**
|