README.md

---
base_model: akhooli/sbert_ar_nli_500k_norm
library_name: setfit
metrics:
- accuracy
pipeline_tag: text-classification
tags:
- setfit
- sentence-transformers
- text-classification
- generated_from_setfit_trainer
widget:
- text: 'بلد مخيف، صار القتل بحجه الشرف متل قتل بعوضة، واللي بيخوف اكتر من اللي واقف
    مكتف ايديه ومش مساعد. وين كنآ، ووين وصلنآ، لمتى حنضل عايشين وساكتين!

    '
- text: "من خلال المتابعة ..يتضح أن أكثر اللاعبين الذين يتم تسويقهم هم لاعبي امريكا\
    \ الجنوبية وأقلهم الافارقة.  \nمن خلال الواقع ..أكثر اللاعبين تهاونا ولعب على\
    \ الواقف  في آخر ٦ شهور من عقودهم  هم لاعبي امريكا الجنوبية ."
- text: ' علم الحزب يا فهمانه ما حطوا لانه عم يحكي وطنيا ومشان ماحدا متلك يعترض. اذا
    حطوا بتعترضي واذا ما حطوا كمان بتعترضي.'
- text: "شيوعي \nعلماني \nمسيحي\nانصار سنه \nصوفي \nيمثلك التجمع \nلا يمثلك التجمع\
    \ \nاهلا بكم جميعا فنحن نريد بناء وطن ❤"
- text: كنا نهرب بحصة الرياضيات والمحاسبة وبنرجع آخر الحصة بخمس دئايئ ولمن تسئلنا
    المعلمة بنحكيلها كنا عند المديرة وبتسدئنا وضلينا
inference: true
model-index:
- name: SetFit with akhooli/sbert_ar_nli_500k_norm
  results:
  - task:
      type: text-classification
      name: Text Classification
    dataset:
      name: Unknown
      type: unknown
      split: test
    metrics:
    - type: accuracy
      value: 0.8606060606060606
      name: Accuracy
---

This is a setfit hate speech detection model (86 % accuracy/f1) based on the [Ar Hate Speech dataset](https://huggingface.co/datasets/akhooli/ar_hs).  

Usage:
```python
pip install setfit
from setfit import SetFitModel
from unicodedata import normalize

# Download model from Hub
model = SetFitModel.from_pretrained("akhooli/setfit_ar_hs")
# Run inference
queries = [
       "سكت دهراً و نطق كفراً",
       "الخلاف ﻻ يفسد للود قضية.",
       "أنت شخص منبوذ. احترم أسيادك.",
       "دع المكارم ﻻ ترحل لبغيتها واقعد فإنك أنت الطاعم الكاسي",
   ]
queries_n = [normalize('NFKC', query) for query in queries]
preds = model.predict(queries_n)
print(preds)
# if you want to see the probabilities for each label
probas = model.predict_proba(queries_n)
print(probas)
```
* [LinkedIn article](https://www.linkedin.com/posts/akhooli_arabic-hate-speech-detection-is-not-an-easy-activity-7261021456023609344-UJzM)
  
The rest of this content is auto-generated.

# SetFit with akhooli/sbert_ar_nli_500k_norm

This is a [SetFit](https://github.com/huggingface/setfit) model that can be used for Text Classification. This SetFit model uses [akhooli/sbert_ar_nli_500k_norm](https://huggingface.co/akhooli/sbert_ar_nli_500k_norm) as the Sentence Transformer embedding model. A [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance is used for classification.

The model has been trained using an efficient few-shot learning technique that involves:

1. Fine-tuning a [Sentence Transformer](https://www.sbert.net) with contrastive learning.
2. Training a classification head with features from the fine-tuned Sentence Transformer.

## Model Details

### Model Description
- **Model Type:** SetFit
- **Sentence Transformer body:** [akhooli/sbert_ar_nli_500k_norm](https://huggingface.co/akhooli/sbert_ar_nli_500k_norm)
- **Classification head:** a [LogisticRegression](https://scikit-learn.org/stable/modules/generated/sklearn.linear_model.LogisticRegression.html) instance
- **Maximum Sequence Length:** 512 tokens
- **Number of Classes:** 2 classes
<!-- - **Training Dataset:** [Unknown](https://huggingface.co/datasets/unknown) -->
<!-- - **Language:** Unknown -->
<!-- - **License:** Unknown -->

### Model Sources

- **Repository:** [SetFit on GitHub](https://github.com/huggingface/setfit)
- **Paper:** [Efficient Few-Shot Learning Without Prompts](https://arxiv.org/abs/2209.11055)
- **Blogpost:** [SetFit: Efficient Few-Shot Learning Without Prompts](https://huggingface.co/blog/setfit)

### Model Labels
| Label    | Examples                                                                                                                                                                                                                                       |
|:---------|:-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|
| positive | <ul><li>' سبحان الله الفلسطينيين شعب خاين في كل مكان \nلاحول ولا قوة إلا بالله'</li><li>'يا بيك عّم تخبرنا عن شي ما فينا تعملو نحن ماًعندنا نواب ولا وزراء بمثلونا بالدولة الا اذا زهقان وعبالك ليك'</li><li>'جوز كذابين منافقين...'</li></ul> |
| negative | <ul><li>'ربي لا تجعلني أسيء الظن بأحد ولا تجعل في قلبي شيئا على أحد ، اللهم أسألك قلباً نقياً صافيا'</li><li>'هشام حداد عامل فيها جون ستيوارت'</li><li>'     بحياة اختك من وين بتجيبي اخبارك؟؟ من صغري وانا عبالي كون... LINK'</li></ul>       |

## Evaluation

### Metrics
| Label   | Accuracy |
|:--------|:---------|
| **all** | 0.8606   |

## Uses

### Direct Use for Inference

First install the SetFit library:

```bash
pip install setfit
```

Then you can load this model and run inference.

```python
from setfit import SetFitModel

# Download from the 🤗 Hub
model = SetFitModel.from_pretrained("akhooli/setfit_ar_hs")
# Run inference
preds = model("شيوعي 
علماني 
مسيحي
انصار سنه 
صوفي 
يمثلك التجمع 
لا يمثلك التجمع 
اهلا بكم جميعا فنحن نريد بناء وطن ❤")
```

<!--
### Downstream Use

*List how someone could finetune this model on their own dataset.*
-->

<!--
### Out-of-Scope Use

*List how the model may foreseeably be misused and address what users ought not to do with the model.*
-->

<!--
## Bias, Risks and Limitations

*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
-->

<!--
### Recommendations

*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
-->

## Training Details

### Training Set Metrics
| Training set | Min | Median  | Max |
|:-------------|:----|:--------|:----|
| Word count   | 1   | 18.8448 | 185 |

| Label    | Training Sample Count |
|:---------|:----------------------|
| negative | 5200                  |
| positive | 4943                  |

### Training Hyperparameters
- batch_size: (32, 32)
- num_epochs: (1, 1)
- max_steps: 6000
- sampling_strategy: undersampling
- body_learning_rate: (2e-05, 1e-05)
- head_learning_rate: 0.01
- loss: CosineSimilarityLoss
- distance_metric: cosine_distance
- margin: 0.25
- end_to_end: False
- use_amp: False
- warmup_proportion: 0.1
- l2_weight: 0.01
- seed: 42
- run_name: setfit_hate_52k_aub_6k
- eval_max_steps: -1
- load_best_model_at_end: False

### Training Results
| Epoch  | Step | Training Loss | Validation Loss |
|:------:|:----:|:-------------:|:---------------:|
| 0.0003 | 1    | 0.3151        | -               |
| 0.0333 | 100  | 0.2902        | -               |
| 0.0667 | 200  | 0.248         | -               |
| 0.1    | 300  | 0.2011        | -               |
| 0.1333 | 400  | 0.164         | -               |
| 0.1667 | 500  | 0.136         | -               |
| 0.2    | 600  | 0.1162        | -               |
| 0.2333 | 700  | 0.0915        | -               |
| 0.2667 | 800  | 0.0724        | -               |
| 0.3    | 900  | 0.0656        | -               |
| 0.3333 | 1000 | 0.05          | -               |
| 0.3667 | 1100 | 0.0454        | -               |
| 0.4    | 1200 | 0.0407        | -               |
| 0.4333 | 1300 | 0.0318        | -               |
| 0.4667 | 1400 | 0.0338        | -               |
| 0.5    | 1500 | 0.0289        | -               |
| 0.5333 | 1600 | 0.0266        | -               |
| 0.5667 | 1700 | 0.0238        | -               |
| 0.6    | 1800 | 0.02          | -               |
| 0.6333 | 1900 | 0.0167        | -               |
| 0.6667 | 2000 | 0.0168        | -               |
| 0.7    | 2100 | 0.0161        | -               |
| 0.7333 | 2200 | 0.0143        | -               |
| 0.7667 | 2300 | 0.0128        | -               |
| 0.8    | 2400 | 0.0128        | -               |
| 0.8333 | 2500 | 0.0146        | -               |
| 0.8667 | 2600 | 0.0113        | -               |
| 0.9    | 2700 | 0.0146        | -               |
| 0.9333 | 2800 | 0.0109        | -               |
| 0.9667 | 2900 | 0.0128        | -               |
| 1.0    | 3000 | 0.0101        | -               |
| 1.0333 | 3100 | 0.0126        | -               |
| 1.0667 | 3200 | 0.0092        | -               |
| 1.1    | 3300 | 0.0108        | -               |
| 1.1333 | 3400 | 0.0095        | -               |
| 1.1667 | 3500 | 0.0121        | -               |
| 1.2    | 3600 | 0.0088        | -               |
| 1.2333 | 3700 | 0.0086        | -               |
| 1.2667 | 3800 | 0.0075        | -               |
| 1.3    | 3900 | 0.009         | -               |
| 1.3333 | 4000 | 0.008         | -               |
| 1.3667 | 4100 | 0.0051        | -               |
| 1.4    | 4200 | 0.007         | -               |
| 1.4333 | 4300 | 0.0055        | -               |
| 1.4667 | 4400 | 0.0074        | -               |
| 1.5    | 4500 | 0.0065        | -               |
| 1.5333 | 4600 | 0.0086        | -               |
| 1.5667 | 4700 | 0.0064        | -               |
| 1.6    | 4800 | 0.0064        | -               |
| 1.6333 | 4900 | 0.0073        | -               |
| 1.6667 | 5000 | 0.0052        | -               |
| 1.7    | 5100 | 0.0056        | -               |
| 1.7333 | 5200 | 0.0059        | -               |
| 1.7667 | 5300 | 0.0048        | -               |
| 1.8    | 5400 | 0.0044        | -               |
| 1.8333 | 5500 | 0.003         | -               |
| 1.8667 | 5600 | 0.0045        | -               |
| 1.9    | 5700 | 0.0043        | -               |
| 1.9333 | 5800 | 0.0042        | -               |
| 1.9667 | 5900 | 0.0029        | -               |
| 2.0    | 6000 | 0.0033        | -               |

### Framework Versions
- Python: 3.10.14
- SetFit: 1.2.0.dev0
- Sentence Transformers: 3.3.0
- Transformers: 4.45.1
- PyTorch: 2.4.0
- Datasets: 3.0.1
- Tokenizers: 0.20.0

## Citation

### BibTeX
```bibtex
@article{https://doi.org/10.48550/arxiv.2209.11055,
    doi = {10.48550/ARXIV.2209.11055},
    url = {https://arxiv.org/abs/2209.11055},
    author = {Tunstall, Lewis and Reimers, Nils and Jo, Unso Eun Seo and Bates, Luke and Korat, Daniel and Wasserblat, Moshe and Pereg, Oren},
    keywords = {Computation and Language (cs.CL), FOS: Computer and information sciences, FOS: Computer and information sciences},
    title = {Efficient Few-Shot Learning Without Prompts},
    publisher = {arXiv},
    year = {2022},
    copyright = {Creative Commons Attribution 4.0 International}
}
```

<!--
## Glossary

*Clearly define terms in order to be accessible across audiences.*
-->

<!--
## Model Card Authors

*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
-->

<!--
## Model Card Contact

*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
-->