leewonjun/e5-mul-0910a

SentenceTransformer based on intfloat/multilingual-e5-large

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-large on the train dataset. 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: 256 tokens
• Output Dimensionality: 1024 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • train

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': 256, 'do_lower_case': False, 'architecture': 'XLMRobertaModel'})
  (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("sentence_transformers_model_id")
# Run inference
sentences = [
    'query: 3D-TSV가 반도체 설계에 미치는 영향은 무엇인가요?',
    'passage: 3 Dimension-Through Silicon Via (기술)',
    'passage: Available Bit Rate (Applicational)',
]
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,  0.8098, -0.1741],
#         [ 0.8098,  1.0000, -0.2449],
#         [-0.1741, -0.2449,  1.0000]])

Evaluation

Metrics

Information Retrieval

• Dataset: e5-eval-real
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.9683
cosine_accuracy@3	0.9981
cosine_accuracy@5	0.9997
cosine_accuracy@10	0.9999
cosine_precision@1	0.9683
cosine_precision@3	0.3327
cosine_precision@5	0.1999
cosine_precision@10	0.1
cosine_recall@1	0.9683
cosine_recall@3	0.9981
cosine_recall@5	0.9997
cosine_recall@10	0.9999
cosine_ndcg@10	0.9874
cosine_mrr@10	0.983
cosine_map@100	0.983

Training Details

Training Dataset

train

• Dataset: train
• Size: 43,318 training samples
• Columns: 0 and 1
• Approximate statistics based on the first 1000 samples:

  	0	1
  type	string	string
  details	min: 11 tokens mean: 17.56 tokens max: 32 tokens	min: 9 tokens mean: 14.13 tokens max: 27 tokens

• Samples:

  0	1
  query: ABPL은 ATM의 기초 속도를 지원하는 물리 계층 장치로 어떻게 구성되나요?	passage: ATM Base Rate Physical Layer Unit (기술)
  query: How is the ABPL configured as a physical layer unit supporting the ATM base rate?	passage: ATM Base Rate Physical Layer Unit (Technical)
  query: ABPL의 역할은 ATM 네트워크에서 무엇을 의미하나요?	passage: ATM Base Rate Physical Layer Unit (개념)

• 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: 64
• per_device_eval_batch_size: 64
• learning_rate: 1e-05
• weight_decay: 0.01
• lr_scheduler_type: cosine
• warmup_ratio: 0.1
• bf16: 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: 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
• torch_empty_cache_steps: None
• learning_rate: 1e-05
• weight_decay: 0.01
• 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: cosine
• 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: True
• 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}
• parallelism_config: 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: no_duplicates
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	e5-eval-real_cosine_ndcg@10
0.0015	1	2.8346	-
0.1477	100	1.1145	-
0.2954	200	0.0332	0.9633
0.4431	300	0.0185	-
0.5908	400	0.0154	0.9782
0.7386	500	0.0116	-
0.8863	600	0.0107	0.9810
1.0340	700	0.0078	-
1.1817	800	0.0076	0.9830
1.3294	900	0.0045	-
1.4771	1000	0.0043	0.9851
1.6248	1100	0.0034	-
1.7725	1200	0.0037	0.9862
1.9202	1300	0.0031	-
2.0679	1400	0.0034	0.9870
2.2157	1500	0.0029	-
2.3634	1600	0.0025	0.9872
2.5111	1700	0.0027	-
2.6588	1800	0.0022	0.9875
2.8065	1900	0.0027	-
2.9542	2000	0.0025	0.9875
-1	-1	-	0.9874

Framework Versions

• Python: 3.12.11
• Sentence Transformers: 5.1.0
• Transformers: 4.56.1
• PyTorch: 2.8.0+cu126
• Accelerate: 1.10.1
• Datasets: 3.6.0
• Tokenizers: 0.22.0

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}
}