metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - loss:MultipleNegativesRankingLoss
  - mteb
base_model: NAMAA-Space/AraModernBert-Base-V1.0
widget:
  - source_sentence: الذكاء الاصطناعي يغير طريقة تفاعلنا مع التكنولوجيا.
    sentences:
      - التكنولوجيا تتطور بسرعة بفضل الذكاء الاصطناعي.
      - الذكاء الاصطناعي يسهم في تطوير التطبيقات الذكية.
      - تحديات الذكاء الاصطناعي تشمل الحفاظ على الأمان والأخلاقيات.
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - pearson_cosine
  - spearman_cosine
model-index:
  - name: NAMAA-Space/AraModernBert-Base-V1.0
    results:
      - dataset:
          config: ar-ar
          name: MTEB STS17 (ar-ar)
          revision: faeb762787bd10488a50c8b5be4a3b82e411949c
          split: test
          type: mteb/sts17-crosslingual-sts
        metrics:
          - type: pearson
            value: 82.4888
          - type: spearman
            value: 83.0981
          - type: cosine_pearson
            value: 82.4888
          - type: cosine_spearman
            value: 83.1109
          - type: manhattan_pearson
            value: 81.2051
          - type: manhattan_spearman
            value: 83.0197
          - type: euclidean_pearson
            value: 81.1013
          - type: euclidean_spearman
            value: 82.8922
          - type: main_score
            value: 83.1109
        task:
          type: STS
      - dataset:
          config: ar
          name: MTEB STS22.v2 (ar)
          revision: d31f33a128469b20e357535c39b82fb3c3f6f2bd
          split: test
          type: mteb/sts22-crosslingual-sts
        metrics:
          - type: pearson
            value: 52.58540000000001
          - type: spearman
            value: 61.7371
          - type: cosine_pearson
            value: 52.58540000000001
          - type: cosine_spearman
            value: 61.7371
          - type: manhattan_pearson
            value: 55.887299999999996
          - type: manhattan_spearman
            value: 61.3654
          - type: euclidean_pearson
            value: 55.633500000000005
          - type: euclidean_spearman
            value: 61.2124
          - type: main_score
            value: 61.7371
        task:
          type: STS
license: apache-2.0
language:
  - ar

SentenceTransformer based on NAMAA-Space/AraModernBert-Base-V1.0

This SentenceTransformer is fine-tuned from NAMAA-Space/AraModernBert-Base-V1.0, bringing strong arabic embeddings useful for a multiple of use cases.

🔹 768-dimensional dense vectors 🎯
🔹 Excels in: Semantic Similarity, Search, Paraphrase Mining, Clustering, Text Classification & More!
🔹 Optimized for speed & efficiency without sacrificing performance

Whether you're building intelligent search engines, chatbots, or AI-powered knowledge graphs, this model delivers meaningful representations of Arabic text with precision and depth.

Try it out & bring Arabic NLP to the next level! 🔥✨

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: ModernBertModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("NAMAA-Space/AraModernBert-Base-STS")
# Run inference
sentences = [
    'الذكاء الاصطناعي يغير طريقة تفاعلنا مع التكنولوجيا.',
    'التكنولوجيا تتطور بسرعة بفضل الذكاء الاصطناعي.',
    'الذكاء الاصطناعي يسهم في تطوير التطبيقات الذكية.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Evaluation

Metrics

Semantic Similarity

Datasets: STS17 and STS22.v2
Evaluated with EmbeddingSimilarityEvaluator

Metric	STS17	STS22.v2
pearson_cosine	0.8249	0.5259
spearman_cosine	0.831	0.6169

Framework Versions

Python: 3.10.12
Sentence Transformers: 3.4.1
Transformers: 4.49.0
PyTorch: 2.1.0+cu118
Accelerate: 1.4.0
Datasets: 2.21.0
Tokenizers: 0.21.0

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}