---
datasets:
- oscar-corpus/OSCAR-2201
language:
- ar
- fr
---

# Model Card for ABDUL-Ar_20k

ABDUL-{VARIANT} is a **BERT-base masked language model** pretrained on **phoneme-normalized Modern Standard Arabic (MSA)** and optionally **normalized French** (depending on the variant). It is designed for **North African Arabic dialects (e.g., Algerian, Moroccan Darija)** even though it is trained **only on formal data (MSA + French)**. Variants differ by **vocab size (20k/30k/40k)** and **training mix (Ar | Ar+Fr | Ar+Fr+CS)**.

---

## Model Details

### Model Description

- **Developed by:** [Yassine Toughrai, Kamel Smaili, David Langlois / LORIA]  
- **Funded by [optional]:** [ANR]  
- **Shared by:** [YassineToughrai]  
- **Model type:** BERT encoder (MLM objective only)  
- **Language(s):** Arabic (dialects + MSA), French  
- **License:** Apache 2.0  
- **Finetuned from:** None (trained from scratch)  

### Model Sources

- **Repository:** [[Ar_20k](https://huggingface.co/YassineToughrai/Ar_20k)]  
- **Paper:** *Modeling North African Dialects from Standard Languages* (ArabicNLP 2025)  
 

---

## Uses

### Direct Use
- As a pretrained encoder for **feature extraction** (hidden states, embeddings).  
- Fill-mask experiments on normalized MSA / dialect input.  

### Downstream Use
- Fine-tuning for **NER** (e.g., DzNER, DarNER, WikiFANE).  
- Fine-tuning for **sentiment / polarity classification** (e.g., TwiFil).  
- Other token-level classification tasks where **North African dialects** or **MSA** are involved.  

### Out-of-Scope Use
- Performance drops significantly on **unnormalized raw dialect text** (requires preprocessing).  
- Not evaluated for **text generation, speech, ASR, or diacritized Arabic**.  

---

## Bias, Risks, and Limitations

- **Bias:** Training data is OSCAR web text (Arabic + French), which may contain social, political, or cultural biases.  
- **Risks:** Applying the model without preprocessing can lead to high OOV rates and poor predictions.  
- **Limitations:** Evaluated mainly on NER and sentiment; generalization to other tasks is untested.  

### Recommendations
- Always apply the same **normalization procedure** before tokenization.  
- Evaluate on your target domain before deployment in real-world applications.  

---

## How to Get Started with the Model

Use the code below to get started with the model.

[More Information Needed]

## Training Details

### Training Data

<!-- This should link to a Dataset Card, perhaps with a short stub of information on what the training data is all about as well as documentation related to data pre-processing or additional filtering. -->

[More Information Needed]

### Training Procedure

<!-- This relates heavily to the Technical Specifications. Content here should link to that section when it is relevant to the training procedure. -->

#### Preprocessing [optional]

[More Information Needed]


#### Training Hyperparameters

- **Training regime:** [More Information Needed] <!--fp32, fp16 mixed precision, bf16 mixed precision, bf16 non-mixed precision, fp16 non-mixed precision, fp8 mixed precision -->

#### Speeds, Sizes, Times [optional]

<!-- This section provides information about throughput, start/end time, checkpoint size if relevant, etc. -->

[More Information Needed]

## Evaluation

<!-- This section describes the evaluation protocols and provides the results. -->

### Testing Data, Factors & Metrics

#### Testing Data

<!-- This should link to a Dataset Card if possible. -->

[More Information Needed]

#### Factors

<!-- These are the things the evaluation is disaggregating by, e.g., subpopulations or domains. -->

[More Information Needed]

#### Metrics

<!-- These are the evaluation metrics being used, ideally with a description of why. -->

[More Information Needed]

### Results

[More Information Needed]

#### Summary



## Model Examination [optional]

<!-- Relevant interpretability work for the model goes here -->

[More Information Needed]

## Environmental Impact

<!-- Total emissions (in grams of CO2eq) and additional considerations, such as electricity usage, go here. Edit the suggested text below accordingly -->

Carbon emissions can be estimated using the [Machine Learning Impact calculator](https://mlco2.github.io/impact#compute) presented in [Lacoste et al. (2019)](https://arxiv.org/abs/1910.09700).

- **Hardware Type:** [More Information Needed]
- **Hours used:** [More Information Needed]
- **Cloud Provider:** [More Information Needed]
- **Compute Region:** [More Information Needed]
- **Carbon Emitted:** [More Information Needed]

## Technical Specifications [optional]

### Model Architecture and Objective

[More Information Needed]

### Compute Infrastructure

[More Information Needed]

#### Hardware

[More Information Needed]

#### Software

[More Information Needed]

## Citation [optional]

<!-- If there is a paper or blog post introducing the model, the APA and Bibtex information for that should go in this section. -->

**BibTeX:**

[More Information Needed]

**APA:**

[More Information Needed]

## Glossary [optional]

<!-- If relevant, include terms and calculations in this section that can help readers understand the model or model card. -->

[More Information Needed]

## More Information [optional]

[More Information Needed]

## Model Card Authors [optional]

[More Information Needed]

## Model Card Contact

[More Information Needed]