SentenceTransformer based on intfloat/multilingual-e5-large

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-large. It maps sentences & paragraphs to a 1024-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: intfloat/multilingual-e5-large
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 1024 dimensions
• Similarity Function: Cosine Similarity

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, 'architecture': 'PeftModel'})
  (1): Pooling({'word_embedding_dimension': 1024, '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})
  (2): Normalize()
)

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("kangbeom/multilingual-e5-large")
# Run inference
sentences = [
    'query: 주문의 설명은 무엇인가?',
    'passage: 〈표 2〉 천연비누 쇼핑몰 시스템의 단어 사전 일부 단어영문명약어명동의어설명주문ORDERORD상품의 생산이나 서비스 의 제공을 요구번호NUMBERNo차례를 나타내거나 식별하기 위해 붙이는 숫자일자DATEDT날짜, 일날짜...',
    'passage: 〈표 $22$〉 Binary NAF Method 스칼라곱 연산과 RENAF Method 멱 승 연산 종류연산 결과 $(mathrmX=2)(\; \backslash \backslash mathrm\; \{\; X\; \}\; =2)$BinaryNAF $(7X)(7X)$\\( (100-1)_ {\\text { NAF } } \\cdot 2=2 * 2 \\rightarrow 2 * 4 \\rightarrow 2 * 3-2=14 \\)RENAF $left(X^{7}right)\backslash \backslash left\; (X\; ^\; \{\; 7\; \}\; \backslash \backslash right\; )$\\( 2 ^ { (100)-11_ {\\text { RENAF } } } =2 * 2 \\rightarrow 4 * 4 \\rightarrow 16 * 16 / 2=128 \\)',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.5568, 0.1304],
#         [0.5568, 1.0000, 0.1720],
#         [0.1304, 0.1720, 1.0000]])

Evaluation

Metrics

Information Retrieval

• Dataset: lora-evaluation
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.467
cosine_accuracy@3	0.699
cosine_accuracy@5	0.786
cosine_accuracy@10	0.8755
cosine_precision@1	0.467
cosine_precision@3	0.233
cosine_precision@5	0.1572
cosine_precision@10	0.0876
cosine_recall@1	0.467
cosine_recall@3	0.699
cosine_recall@5	0.786
cosine_recall@10	0.8755
cosine_ndcg@10	0.668
cosine_mrr@10	0.6018
cosine_map@100	0.6076

Training Details

Training Dataset

Unnamed Dataset

• Size: 7,200 training samples
• Columns: sentence_0 and sentence_1
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1
  type	string	string
  details	min: 11 tokens mean: 28.15 tokens max: 105 tokens	min: 10 tokens mean: 127.24 tokens max: 512 tokens

• Samples:

  sentence_0	sentence_1
  query: 실험에서 그림 12(d)는 참값에 아주 근접했나요?	passage: 그림 12(d)는 본 논문에서 제안하는 알고리즘을 이용한 디블러링 결과영상을 이용하여 3차원 형상 복원을 수행한 결과이다.
  query: 표에서, 15mm의 메쉬분할 수행시간은 얼마인가?	passage: 〈표 5〉 임계값에 따른 메쉬정보 및 수행시간(데이터 A) 임계값 ( ( \mathrm { mm } ) )분할전(15 )(10 )(5 )총메쉬수(8,200 )(163,124 )(297,207 )( 1,185,145 )평균에지길이( 20.43 )( 6.52 )( 4.85 )( 2.45 )수행시간 (sec)메쉬분할( 0.113 )( 0.181 )( 0.785 )거리기반 대응( 0.137 )( 2.966 )( 5.453 )( 21.228 )
  query: 기업 시스템 인증 및 정보자산 보호관리에 주로 사용되는 표준은 무엇일까?	passage: 각 표준 특성 및 취약점 표준특성단일 표준으로 적용 시 취약점ISO 20022금융기관 상호 운영을 위한 표준 모듈간 상호보안 부족 클라이언트의 행동에 관한 보안기능 부족ISO 27001기업 시스템 인증 및 정보자산 보호관리 모듈간 상호보안 부족 기술적인 보안 부족Common CriteriaIT 제품의 개발, 평가, 조달 지침현재 인터넷 뱅킹 관련 보호프로파일 없음웹 환경 구축 및 운영을 위한 보안 관리 지침웹 환경 안전을 위한 기술 특화된 위협에 대한 대응 부족 기능 요구사항 부족

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 4
• per_device_eval_batch_size: 4
• num_train_epochs: 2
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 4
• per_device_eval_batch_size: 4
• 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: 5e-05
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1
• num_train_epochs: 2
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.0
• 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: False
• 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_fused
• 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: None
• hub_always_push: False
• hub_revision: None
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• include_for_metrics: []
• 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
• 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
• use_liger_kernel: False
• liger_kernel_config: None
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: round_robin
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	lora-evaluation_cosine_ndcg@10
0.2778	500	0.2637	0.6548
0.5556	1000	0.0663	0.6597
0.8333	1500	0.0647	0.6609
1.0	1800	-	0.6658
1.1111	2000	0.0432	0.6680

Framework Versions

• Python: 3.12.11
• Sentence Transformers: 5.1.0
• Transformers: 4.55.4
• PyTorch: 2.8.0+cu126
• Accelerate: 1.10.1
• Datasets: 4.0.0
• Tokenizers: 0.21.4

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}
}