akhooli/sbert_ar_nli_500k_norm

BERT base trained on 500k Arabic NLI triplets

This is a sentence-transformers model finetuned from aubmindlab/bert-base-arabertv02. 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. Most of the contents of this card are auto-generated. The dataset used was normalized before training, so it is recommended to normalize your queries and documents:

from unicodedata import normalize
query_n = normalize('NFKC', query)

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: aubmindlab/bert-base-arabertv02
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity
• Language: ar
• License: apache-2.0

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: BertModel 
  (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("sentence_transformers_model_id")
# Run inference
sentences = [
    'لماذا رائحة الغسيل بعد الغسيل',
    'كيفية التخلص من الرائحة: إذا تم اكتشافها على الفور ، فعادة ما يكون غسل الملابس مرة أخرى هو كل ما يلزم لإزالة الروائح الكريهة. إذا لم ينجح ذلك وما زلت تواجه مشاكل ، جرب إحدى الطرق التالية: اغسل مرة أخرى ولكن هذه المرة أضف كوبًا واحدًا من الخل إلى الحمولة (جنبًا إلى جنب مع منظف الغسيل).',
    'حافظ على ملابسك منتعشة وجافة بالمجفف الكهربائي من سيرز. عندما يدور يوم الغسيل ، يمكنك الاعتماد على الأداء الفعال للمجفف الكهربائي. يحتوي سيرز على مجففات تناسب ديكور أي غرفة غسيل. من الفولاذ المقاوم للصدأ إلى تشطيبات الأونيكس ، يسهل تنسيق هذا الجهاز الأنيق مع الغسالة التي تختارها.',
]
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]

Training Details

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 32
• learning_rate: 2e-05
• num_train_epochs: 1
• warmup_ratio: 0.1
• fp16: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 32
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 2e-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: 1
• 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
• restore_callback_states_from_checkpoint: 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, 'non_blocking': False, '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_eval_metrics: False
• eval_on_start: False
• eval_use_gather_object: False
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	loss
0.016	250	4.065	-
0.032	500	1.9888	-
0.048	750	1.4433	-
0.064	1000	1.2331	-
0.08	1250	1.0376	-
0.096	1500	1.0282	-
0.112	1750	0.892	-
0.128	2000	0.8597	-
0.144	2250	0.8105	-
0.16	2500	0.8397	-
0.176	2750	0.7831	-
0.192	3000	0.7718	-
0.208	3250	0.7466	-
0.224	3500	0.7199	-
0.24	3750	0.717	-
0.256	4000	0.6693	-
0.272	4250	0.6473	-
0.288	4500	0.6388	-
0.304	4750	0.6226	-
0.32	5000	0.5967	0.5908
0.336	5250	0.5971	-
0.352	5500	0.6453	-
0.368	5750	0.6468	-
0.384	6000	0.5849	-
0.4	6250	0.5552	-
0.416	6500	0.512	-
0.432	6750	0.5825	-
0.448	7000	0.5895	-
0.464	7250	0.5453	-
0.48	7500	0.5453	-
0.496	7750	0.5034	-
0.512	8000	0.5756	-
0.528	8250	0.5576	-
0.544	8500	0.55	-
0.56	8750	0.5191	-
0.576	9000	0.5104	-
0.592	9250	0.5062	-
0.608	9500	0.5493	-
0.624	9750	0.5131	-
0.64	10000	0.5276	0.4551
0.656	10250	0.5332	-
0.672	10500	0.5076	-
0.688	10750	0.511	-
0.704	11000	0.448	-
0.72	11250	0.4928	-
0.736	11500	0.4492	-
0.752	11750	0.4514	-
0.768	12000	0.4749	-
0.784	12250	0.4765	-
0.8	12500	0.5043	-
0.816	12750	0.4617	-
0.832	13000	0.4845	-
0.848	13250	0.5101	-
0.864	13500	0.4759	-
0.88	13750	0.4824	-
0.896	14000	0.4646	-
0.912	14250	0.4504	-
0.928	14500	0.4603	-
0.944	14750	0.4645	-
0.96	15000	0.4688	0.4054
0.976	15250	0.506	-
0.992	15500	0.4604	-

Framework Versions

• Python: 3.10.14
• Sentence Transformers: 3.1.1
• Transformers: 4.44.2
• PyTorch: 2.4.0
• Accelerate: 0.34.2
• Datasets: 3.0.0
• 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",
}

Matryoshka2dLoss

@misc{li20242d,
    title={2D Matryoshka Sentence Embeddings},
    author={Xianming Li and Zongxi Li and Jing Li and Haoran Xie and Qing Li},
    year={2024},
    eprint={2402.14776},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

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}
}