Omartificial-Intelligence-Space
/

Arabic-MiniLM-L12-v2-all-nli-triplet

Feature Extraction

sentence-transformers

sentence-similarity

Generated from Trainer

dataset_size:557850

loss:MatryoshkaLoss

loss:MultipleNegativesRankingLoss

Eval Results (legacy)

text-embeddings-inference

Model card Files Files and versions

SentenceTransformer based on sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2

This is a sentence-transformers model finetuned from sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 on the Omartificial-Intelligence-Space/arabic-n_li-triplet dataset. It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more. This model is part of the Arabic Matryoshka Embedding Models collection. It was presented in the paper GATE: General Arabic Text Embedding for Enhanced Semantic Textual Similarity with Matryoshka Representation Learning and Hybrid Loss Training.

Model Details

Model Description

Model Type: Sentence Transformer
Base model: sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
Maximum Sequence Length: 128 tokens
Output Dimensionality: 384 tokens
Similarity Function: Cosine Similarity
Training Dataset:
- Omartificial-Intelligence-Space/arabic-n_li-triplet

Model Sources

Documentation: Sentence Transformers Documentation
Repository: Sentence Transformers on GitHub
**Hugging

Downloads last month: 132

Safetensors

Model size

0.1B params

Tensor type

F32

·

Model tree for Omartificial-Intelligence-Space/Arabic-MiniLM-L12-v2-all-nli-triplet

Base model

sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2

Finetuned

(298)

this model

Dataset used to train Omartificial-Intelligence-Space/Arabic-MiniLM-L12-v2-all-nli-triplet

Spaces using Omartificial-Intelligence-Space/Arabic-MiniLM-L12-v2-all-nli-triplet 10

Collection including Omartificial-Intelligence-Space/Arabic-MiniLM-L12-v2-all-nli-triplet

Arabic Matryoshka & GATE Embedding Models

A collection of advanced Arabic Matryoshka Embedding Models designed for efficient and high-performance Arabic NLP, available publicly on Hugging Face • 12 items • Updated Nov 28, 2025 • 16

Paper for Omartificial-Intelligence-Space/Arabic-MiniLM-L12-v2-all-nli-triplet

GATE: General Arabic Text Embedding for Enhanced Semantic Textual Similarity with Matryoshka Representation Learning and Hybrid Loss Training

Paper • 2505.24581 • Published May 30, 2025 • 2

Evaluation results

main_score on MTEB MintakaRetrieval (ar)
test set self-reported

12.493
map_at_1 on MTEB MintakaRetrieval (ar)
test set self-reported

5.719
map_at_3 on MTEB MintakaRetrieval (ar)
test set self-reported

8.269
map_at_5 on MTEB MintakaRetrieval (ar)
test set self-reported

9.172
map_at_10 on MTEB MintakaRetrieval (ar)
test set self-reported

9.894
ndcg_at_1 on MTEB MintakaRetrieval (ar)
test set self-reported

5.719
ndcg_at_3 on MTEB MintakaRetrieval (ar)
test set self-reported

9.128
ndcg_at_5 on MTEB MintakaRetrieval (ar)
test set self-reported

10.745
ndcg_at_10 on MTEB MintakaRetrieval (ar)
test set self-reported

12.493
recall_at_1 on MTEB MintakaRetrieval (ar)
test set self-reported

5.719
recall_at_3 on MTEB MintakaRetrieval (ar)
test set self-reported

11.621
recall_at_5 on MTEB MintakaRetrieval (ar)
test set self-reported

15.524
recall_at_10 on MTEB MintakaRetrieval (ar)
test set self-reported

20.926
precision_at_1 on MTEB MintakaRetrieval (ar)
test set self-reported

5.719
precision_at_3 on MTEB MintakaRetrieval (ar)
test set self-reported

3.874
precision_at_5 on MTEB MintakaRetrieval (ar)
test set self-reported

3.105
precision_at_10 on MTEB MintakaRetrieval (ar)
test set self-reported

2.093
mrr_at_1 on MTEB MintakaRetrieval (ar)
test set self-reported

5.720
mrr_at_3 on MTEB MintakaRetrieval (ar)
test set self-reported

8.269
mrr_at_5 on MTEB MintakaRetrieval (ar)
test set self-reported

9.172