Arabic Reranker Model

This is an Arabic reranker model, fine-tuned from the Omartificial-Intelligence-Space/Arabic-Triplet-Matryoshka-V2, which itself is based on aubmindlab/bert-base-arabertv02. The model is designed to perform reranking tasks by scoring and ordering text options based on their relevance to a given query, specifically optimized for Arabic text.

This model was trained on a synthetic dataset of Arabic triplets generated using large language models (LLMs). It was refined using a scoring technique, making it ideal for ranking tasks in Arabic Natural Language Processing (NLP).

Model Use

This model is well-suited for Arabic text reranking tasks, including:

Information retrieval and document ranking

Search engine results reranking

Question-answering tasks requiring ranked answer choices

Example Usage

Below is an example of how to use the model with the sentence_transformers library to rerank paragraphs based on relevance to a query.

Code Example

`python from sentence_transformers import CrossEncoder

Load the model

model = CrossEncoder('oddadmix/arabic-reranker', max_length=512)

Define the query and candidate paragraphs

Query = 'كيف يمكن استخدام التعلم العميق في معالجة الصور الطبية؟' Paragraph1 = 'التعلم العميق يساعد في تحليل الصور الطبية وتشخيص الأمراض' Paragraph2 = 'الذكاء الاصطناعي يستخدم في تحسين الإنتاجية في الصناعات'

Score the paragraphs based on relevance to the query

scores = model.predict([(Query, Paragraph1), (Query, Paragraph2)])

Output scores

print("Score for Paragraph 1:", scores[0]) print("Score for Paragraph 2:", scores[1])