Omartificial-Intelligence-Space/Arabic-mpnet-base-all-nli-triplet

SentenceTransformer based on tomaarsen/mpnet-base-all-nli-triplet

This is a sentence-transformers model finetuned from tomaarsen/mpnet-base-all-nli-triplet on the Omartificial-Intelligence-Space/arabic-n_li-triplet dataset. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: tomaarsen/mpnet-base-all-nli-triplet
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 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 Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: MPNetModel 
  (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("Omartificial-Intelligence-Space/mpnet-base-all-nli-triplet-Arabic-mpnet_base")
# Run inference
sentences = [
    'يجلس شاب ذو شعر أشقر على الحائط يقرأ جريدة بينما تمر امرأة وفتاة شابة.',
    'ذكر شاب ينظر إلى جريدة بينما تمر إمرأتان بجانبه',
    'الشاب نائم بينما الأم تقود ابنتها إلى الحديقة',
]
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

• Dataset: sts-test-768
• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.6699
spearman_cosine	0.6757
pearson_manhattan	0.6943
spearman_manhattan	0.684
pearson_euclidean	0.6973
spearman_euclidean	0.6873
pearson_dot	0.5534
spearman_dot	0.5422
pearson_max	0.6973
spearman_max	0.6873

Semantic Similarity

• Dataset: sts-test-512
• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.6628
spearman_cosine	0.6703
pearson_manhattan	0.6917
spearman_manhattan	0.6816
pearson_euclidean	0.6949
spearman_euclidean	0.6853
pearson_dot	0.5229
spearman_dot	0.5114
pearson_max	0.6949
spearman_max	0.6853

Semantic Similarity

• Dataset: sts-test-256
• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.6368
spearman_cosine	0.6513
pearson_manhattan	0.6832
spearman_manhattan	0.6746
pearson_euclidean	0.6844
spearman_euclidean	0.676
pearson_dot	0.4266
spearman_dot	0.4179
pearson_max	0.6844
spearman_max	0.676

Semantic Similarity

• Dataset: sts-test-128
• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.6148
spearman_cosine	0.6355
pearson_manhattan	0.6731
spearman_manhattan	0.6653
pearson_euclidean	0.6764
spearman_euclidean	0.6691
pearson_dot	0.3513
spearman_dot	0.3445
pearson_max	0.6764
spearman_max	0.6691

Semantic Similarity

• Dataset: sts-test-64
• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.5789
spearman_cosine	0.6081
pearson_manhattan	0.6579
spearman_manhattan	0.6519
pearson_euclidean	0.663
spearman_euclidean	0.6571
pearson_dot	0.2403
spearman_dot	0.2331
pearson_max	0.663
spearman_max	0.6571

Training Details

Training Dataset

Omartificial-Intelligence-Space/arabic-n_li-triplet

• Dataset: Omartificial-Intelligence-Space/arabic-n_li-triplet
• Size: 557,850 training samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 12 tokens mean: 23.93 tokens max: 155 tokens	min: 9 tokens mean: 29.62 tokens max: 117 tokens	min: 13 tokens mean: 33.95 tokens max: 149 tokens

• Samples:

  anchor	positive	negative
  شخص على حصان يقفز فوق طائرة معطلة	شخص في الهواء الطلق، على حصان.	شخص في مطعم، يطلب عجة.
  أطفال يبتسمون و يلوحون للكاميرا	هناك أطفال حاضرون	الاطفال يتجهمون
  صبي يقفز على لوح التزلج في منتصف الجسر الأحمر.	الفتى يقوم بخدعة التزلج	الصبي يتزلج على الرصيف

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          128,
          64
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Evaluation Dataset

Omartificial-Intelligence-Space/arabic-n_li-triplet

• Dataset: Omartificial-Intelligence-Space/arabic-n_li-triplet
• Size: 6,584 evaluation samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 12 tokens mean: 49.5 tokens max: 246 tokens	min: 8 tokens mean: 23.66 tokens max: 103 tokens	min: 9 tokens mean: 25.33 tokens max: 82 tokens

• Samples:

  anchor	positive	negative
  امرأتان يتعانقان بينما يحملان حزمة	إمرأتان يحملان حزمة	الرجال يتشاجرون خارج مطعم
  طفلين صغيرين يرتديان قميصاً أزرق، أحدهما يرتدي الرقم 9 والآخر يرتدي الرقم 2 يقفان على خطوات خشبية في الحمام ويغسلان أيديهما في المغسلة.	طفلين يرتديان قميصاً مرقماً يغسلون أيديهم	طفلين يرتديان سترة يذهبان إلى المدرسة
  رجل يبيع الدونات لعميل خلال معرض عالمي أقيم في مدينة أنجليس	رجل يبيع الدونات لعميل	امرأة تشرب قهوتها في مقهى صغير

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          128,
          64
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• warmup_ratio: 0.1
• fp16: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• prediction_loss_only: True
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• learning_rate: 5e-05
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 3
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.1
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• save_safetensors: True
• save_on_each_node: False
• save_only_model: False
• no_cuda: False
• use_cpu: False
• use_mps_device: False
• seed: 42
• data_seed: None
• jit_mode_eval: False
• use_ipex: False
• bf16: False
• fp16: True
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: 0
• ddp_backend: None
• tpu_num_cores: None
• tpu_metrics_debug: False
• debug: []
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• past_index: -1
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: []
• fsdp_min_num_params: 0
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• fsdp_transformer_layer_cls_to_wrap: None
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'gradient_accumulation_kwargs': None}
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch
• optim_args: None
• adafactor: False
• group_by_length: False
• length_column_name: length
• ddp_find_unused_parameters: None
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• dataloader_pin_memory: True
• dataloader_persistent_workers: False
• skip_memory_metrics: True
• use_legacy_prediction_loop: False
• push_to_hub: False
• resume_from_checkpoint: None
• hub_model_id: None
• hub_strategy: every_save
• hub_private_repo: False
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• eval_do_concat_batches: True
• fp16_backend: auto
• push_to_hub_model_id: None
• push_to_hub_organization: None
• mp_parameters:
• auto_find_batch_size: False
• full_determinism: False
• torchdynamo: None
• ray_scope: last
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• dispatch_batches: None
• split_batches: None
• include_tokens_per_second: False
• include_num_input_tokens_seen: False
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Click to expand

Epoch	Step	Training Loss	sts-test-128_spearman_cosine	sts-test-256_spearman_cosine	sts-test-512_spearman_cosine	sts-test-64_spearman_cosine	sts-test-768_spearman_cosine
0.0229	200	21.5318	-	-	-	-	-
0.0459	400	17.2344	-	-	-	-	-
0.0688	600	15.393	-	-	-	-	-
0.0918	800	13.7897	-	-	-	-	-
0.1147	1000	13.534	-	-	-	-	-
0.1377	1200	12.2683	-	-	-	-	-
0.1606	1400	10.9271	-	-	-	-	-
0.1835	1600	11.071	-	-	-	-	-
0.2065	1800	10.0153	-	-	-	-	-
0.2294	2000	9.8463	-	-	-	-	-
0.2524	2200	10.0194	-	-	-	-	-
0.2753	2400	9.8371	-	-	-	-	-
0.2983	2600	9.6315	-	-	-	-	-
0.3212	2800	8.9858	-	-	-	-	-
0.3442	3000	9.1876	-	-	-	-	-
0.3671	3200	8.8028	-	-	-	-	-
0.3900	3400	8.6075	-	-	-	-	-
0.4130	3600	8.4285	-	-	-	-	-
0.4359	3800	8.1258	-	-	-	-	-
0.4589	4000	8.2508	-	-	-	-	-
0.4818	4200	7.8037	-	-	-	-	-
0.5048	4400	7.7133	-	-	-	-	-
0.5277	4600	7.5006	-	-	-	-	-
0.5506	4800	7.7025	-	-	-	-	-
0.5736	5000	7.7593	-	-	-	-	-
0.5965	5200	7.6305	-	-	-	-	-
0.6195	5400	7.7502	-	-	-	-	-
0.6424	5600	7.5624	-	-	-	-	-
0.6654	5800	7.5287	-	-	-	-	-
0.6883	6000	7.4261	-	-	-	-	-
0.7113	6200	7.239	-	-	-	-	-
0.7342	6400	7.1631	-	-	-	-	-
0.7571	6600	7.6865	-	-	-	-	-
0.7801	6800	7.6124	-	-	-	-	-
0.8030	7000	6.9936	-	-	-	-	-
0.8260	7200	6.7331	-	-	-	-	-
0.8489	7400	6.4542	-	-	-	-	-
0.8719	7600	6.1994	-	-	-	-	-
0.8948	7800	5.9798	-	-	-	-	-
0.9177	8000	5.7808	-	-	-	-	-
0.9407	8200	5.6952	-	-	-	-	-
0.9636	8400	5.5082	-	-	-	-	-
0.9866	8600	5.4421	-	-	-	-	-
1.0095	8800	3.0309	-	-	-	-	-
1.0026	9000	1.1835	-	-	-	-	-
1.0256	9200	8.1196	-	-	-	-	-
1.0485	9400	8.0326	-	-	-	-	-
1.0715	9600	8.5028	-	-	-	-	-
1.0944	9800	7.6923	-	-	-	-	-
1.1174	10000	8.029	-	-	-	-	-
1.1403	10200	7.5052	-	-	-	-	-
1.1632	10400	7.1177	-	-	-	-	-
1.1862	10600	6.9594	-	-	-	-	-
1.2091	10800	6.6662	-	-	-	-	-
1.2321	11000	6.6903	-	-	-	-	-
1.2550	11200	6.9523	-	-	-	-	-
1.2780	11400	6.676	-	-	-	-	-
1.3009	11600	6.7141	-	-	-	-	-
1.3238	11800	6.568	-	-	-	-	-
1.3468	12000	6.8938	-	-	-	-	-
1.3697	12200	6.3745	-	-	-	-	-
1.3927	12400	6.2513	-	-	-	-	-
1.4156	12600	6.2589	-	-	-	-	-
1.4386	12800	6.1388	-	-	-	-	-
1.4615	13000	6.1835	-	-	-	-	-
1.4845	13200	5.9004	-	-	-	-	-
1.5074	13400	5.7891	-	-	-	-	-
1.5303	13600	5.6184	-	-	-	-	-
1.5533	13800	5.9762	-	-	-	-	-
1.5762	14000	5.9737	-	-	-	-	-
1.5992	14200	5.8563	-	-	-	-	-
1.6221	14400	5.8904	-	-	-	-	-
1.6451	14600	5.8484	-	-	-	-	-
1.6680	14800	5.8906	-	-	-	-	-
1.6909	15000	5.7613	-	-	-	-	-
1.7139	15200	5.5744	-	-	-	-	-
1.7368	15400	5.6569	-	-	-	-	-
1.7598	15600	5.7439	-	-	-	-	-
1.7827	15800	5.5593	-	-	-	-	-
1.8057	16000	5.2935	-	-	-	-	-
1.8286	16200	5.088	-	-	-	-	-
1.8516	16400	5.0167	-	-	-	-	-
1.8745	16600	4.84	-	-	-	-	-
1.8974	16800	4.6731	-	-	-	-	-
1.9204	17000	4.6404	-	-	-	-	-
1.9433	17200	4.6413	-	-	-	-	-
1.9663	17400	4.4495	-	-	-	-	-
1.9892	17600	4.4262	-	-	-	-	-
2.0122	17800	2.01	-	-	-	-	-
2.0053	18000	1.8418	-	-	-	-	-
2.0282	18200	6.2714	-	-	-	-	-
2.0512	18400	6.1742	-	-	-	-	-
2.0741	18600	6.5996	-	-	-	-	-
2.0971	18800	6.0907	-	-	-	-	-
2.1200	19000	6.2418	-	-	-	-	-
2.1429	19200	5.7817	-	-	-	-	-
2.1659	19400	5.7073	-	-	-	-	-
2.1888	19600	5.2645	-	-	-	-	-
2.2118	19800	5.3451	-	-	-	-	-
2.2347	20000	5.2453	-	-	-	-	-
2.2577	20200	5.6161	-	-	-	-	-
2.2806	20400	5.2289	-	-	-	-	-
2.3035	20600	5.3888	-	-	-	-	-
2.3265	20800	5.2483	-	-	-	-	-
2.3494	21000	5.5791	-	-	-	-	-
2.3724	21200	5.1643	-	-	-	-	-
2.3953	21400	5.1231	-	-	-	-	-
2.4183	21600	5.1055	-	-	-	-	-
2.4412	21800	5.1778	-	-	-	-	-
2.4642	22000	5.0466	-	-	-	-	-
2.4871	22200	4.8321	-	-	-	-	-
2.5100	22400	4.7056	-	-	-	-	-
2.5330	22600	4.6858	-	-	-	-	-
2.5559	22800	4.9189	-	-	-	-	-
2.5789	23000	4.912	-	-	-	-	-
2.6018	23200	4.8289	-	-	-	-	-
2.6248	23400	4.8959	-	-	-	-	-
2.6477	23600	4.9441	-	-	-	-	-
2.6706	23800	4.9334	-	-	-	-	-
2.6936	24000	4.8328	-	-	-	-	-
2.7165	24200	4.601	-	-	-	-	-
2.7395	24400	4.834	-	-	-	-	-
2.7624	24600	5.152	-	-	-	-	-
2.7854	24800	4.9232	-	-	-	-	-
2.8083	25000	4.6556	-	-	-	-	-
2.8312	25200	4.6229	-	-	-	-	-
2.8542	25400	4.5768	-	-	-	-	-
2.8771	25600	4.3619	-	-	-	-	-
2.9001	25800	4.3608	-	-	-	-	-
2.9230	26000	4.2834	-	-	-	-	-
2.9403	26151	-	0.6355	0.6513	0.6703	0.6081	0.6757

Framework Versions

• Python: 3.9.18
• Sentence Transformers: 3.0.1
• Transformers: 4.40.0
• PyTorch: 2.2.2+cu121
• Accelerate: 0.26.1
• Datasets: 2.19.0
• Tokenizers: 0.19.1

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",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning}, 
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

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}
}

Acknowledgments

The author would like to thank Prince Sultan University for their invaluable support in this project. Their contributions and resources have been instrumental in the development and fine-tuning of these models.

## Citation

If you use the Arabic Matryoshka Embeddings Model, please cite it as follows:

@misc{nacar2024enhancingsemanticsimilarityunderstanding,
      title={Enhancing Semantic Similarity Understanding in Arabic NLP with Nested Embedding Learning}, 
      author={Omer Nacar and Anis Koubaa},
      year={2024},
      eprint={2407.21139},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2407.21139}, 
}