|
--- |
|
language: |
|
- ms |
|
library_name: transformers |
|
--- |
|
|
|
Safe for Work Classifier Model for Malaysian Data(WORK IN PROGRESS) |
|
|
|
|
|
Finetuned https://huggingface.co/mesolitica/malaysian-mistral-191M-MLM-512 with Malaysian NSFW data. |
|
|
|
|
|
![Image in a markdown cell](https://github.com/mesolitica/malaysian-llmops/raw/main/e2e.png) |
|
|
|
|
|
|
|
### How to use |
|
|
|
```python |
|
from classifier import MistralForSequenceClassification |
|
model = MistralForSequenceClassification.from_pretrained('malaysia-ai/malaysian-sfw-classifier') |
|
``` |
|
|
|
|
|
``` |
|
precision recall f1-score support |
|
|
|
racist 0.96737 0.92272 0.94452 1928 |
|
religion insult 0.94210 0.91785 0.92982 3457 |
|
psychiatric or mental illness 0.97140 0.92858 0.94951 5559 |
|
sexist 0.93154 0.83968 0.88323 1734 |
|
harassment 0.76633 0.94214 0.84519 2005 |
|
informative 0.84307 0.88797 0.86494 1821 |
|
safe for work 0.86179 0.91640 0.88826 1878 |
|
hate 0.87780 0.84726 0.86226 2357 |
|
|
|
accuracy 0.90622 20739 |
|
macro avg 0.89518 0.90033 0.89596 20739 |
|
weighted avg 0.91115 0.90622 0.90725 20739 |
|
``` |