Update README.MD
Browse files
README.md
CHANGED
|
@@ -11,30 +11,29 @@ model-index:
|
|
| 11 |
- name: v1
|
| 12 |
results:
|
| 13 |
- task:
|
| 14 |
-
type:
|
| 15 |
-
name:
|
| 16 |
metrics:
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
|
| 22 |
-
|
| 23 |
-
|
| 24 |
-
|
| 25 |
-
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
|
| 31 |
-
|
| 32 |
-
|
|
|
|
|
|
|
| 33 |
---
|
| 34 |
|
| 35 |
-
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
|
| 36 |
-
should probably proofread and complete it, then remove this comment. -->
|
| 37 |
-
|
| 38 |
# swahBERT
|
| 39 |
This model was trained from scratch on an unknown dataset.
|
| 40 |
It achieves the following results on the evaluation set:
|
|
@@ -44,12 +43,11 @@ It achieves the following results on the evaluation set:
|
|
| 44 |
- Recall: 0.9010
|
| 45 |
- F1: 0.9272
|
| 46 |
## Model description
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
More information needed
|
| 50 |
## Training and evaluation data
|
| 51 |
-
|
| 52 |
-
|
| 53 |
### Training hyperparameters
|
| 54 |
The following hyperparameters were used during training:
|
| 55 |
- learning_rate: 5e-05
|
|
@@ -80,4 +78,17 @@ The following hyperparameters were used during training:
|
|
| 80 |
- Transformers 4.33.1
|
| 81 |
- Pytorch 2.0.1+cpu
|
| 82 |
- Datasets 2.14.5
|
| 83 |
-
- Tokenizers 0.13.3
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 11 |
- name: v1
|
| 12 |
results:
|
| 13 |
- task:
|
| 14 |
+
type: Offensive words classifier
|
| 15 |
+
name: Text Classification
|
| 16 |
metrics:
|
| 17 |
+
- type: f1
|
| 18 |
+
value: 0.9272349272349272
|
| 19 |
+
name: F1 Score
|
| 20 |
+
verified: false
|
| 21 |
+
- type: precision
|
| 22 |
+
value: 0.9550321199143469
|
| 23 |
+
name: Precision
|
| 24 |
+
verified: false
|
| 25 |
+
- type: recall
|
| 26 |
+
value: 0.901010101010101
|
| 27 |
+
name: Recall
|
| 28 |
+
verified: false
|
| 29 |
+
- type: accuracy
|
| 30 |
+
value: 0.9292214357937311
|
| 31 |
+
name: Accuracy
|
| 32 |
+
verified: false
|
| 33 |
+
datasets:
|
| 34 |
+
- metabloit/offensive-swahili-text
|
| 35 |
---
|
| 36 |
|
|
|
|
|
|
|
|
|
|
| 37 |
# swahBERT
|
| 38 |
This model was trained from scratch on an unknown dataset.
|
| 39 |
It achieves the following results on the evaluation set:
|
|
|
|
| 43 |
- Recall: 0.9010
|
| 44 |
- F1: 0.9272
|
| 45 |
## Model description
|
| 46 |
+
This is a fine tuned swahBERT model. You can get the original model from [here](https://github.com/gatimartin/SwahBERT "swahBERT Model")
|
| 47 |
+
|
|
|
|
| 48 |
## Training and evaluation data
|
| 49 |
+
The model was fine tuned using [this dataset](https://huggingface.co/datasets/metabloit/offensive-swahili-text "Swahili offensive/non-offensive dataset")
|
| 50 |
+
|
| 51 |
### Training hyperparameters
|
| 52 |
The following hyperparameters were used during training:
|
| 53 |
- learning_rate: 5e-05
|
|
|
|
| 78 |
- Transformers 4.33.1
|
| 79 |
- Pytorch 2.0.1+cpu
|
| 80 |
- Datasets 2.14.5
|
| 81 |
+
- Tokenizers 0.13.3
|
| 82 |
+
|
| 83 |
+
## References
|
| 84 |
+
@inproceedings{martin-etal-2022-swahbert,
|
| 85 |
+
title = "{S}wah{BERT}: Language Model of {S}wahili",
|
| 86 |
+
author = "Martin, Gati and Mswahili, Medard Edmund and Jeong, Young-Seob and Woo, Jiyoung",
|
| 87 |
+
booktitle = "Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",
|
| 88 |
+
month = jul,
|
| 89 |
+
year = "2022",
|
| 90 |
+
address = "Seattle, United States",
|
| 91 |
+
publisher = "Association for Computational Linguistics",
|
| 92 |
+
url = "https://aclanthology.org/2022.naacl-main.23",
|
| 93 |
+
pages = "303--313"
|
| 94 |
+
}
|