snunlp/KR-SBERT-Medium-extended-klueNLItriplet_PARpair_QApair-klueSTS

SentenceTransformer

This is a sentence-transformers model trained on the klue/klue 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
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity
• Training Dataset:
  • klue/klue
• Language: ko

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("snunlp/KR-SBERT-Medium-extended-klueNLItriplet_PARpair_QApair-klueSTS")
# Run inference
sentences = [
    'SR은 동대구·김천구미·신경주역에서 승하차하는 모든 국민에게 운임 10%를 할인해 준다.',
    'SR은 동대구역, 김천구미역, 신주역을 오가는 모든 승객을 대상으로 요금을 10% 할인해 드립니다.',
    '수강신청 하는 날짜가 어느 날짜인지 아시는지요?',
]
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-dev
• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.8786
spearman_cosine	0.8765
pearson_manhattan	0.8589
spearman_manhattan	0.8582
pearson_euclidean	0.8595
spearman_euclidean	0.8597
pearson_dot	0.8518
spearman_dot	0.8479
pearson_max	0.8786
spearman_max	0.8765

Training Details

Training Dataset

klue/klue

• Dataset: klue/klue at 349481e
• Size: 11,668 training samples
• Columns: sentence1, sentence2, and label
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	label
  type	string	string	float
  details	min: 7 tokens mean: 18.12 tokens max: 56 tokens	min: 6 tokens mean: 17.58 tokens max: 60 tokens	min: 0.0 mean: 0.44 max: 1.0

• Samples:

  sentence1	sentence2	label
  숙소 위치는 찾기 쉽고 일반적인 한국의 반지하 숙소입니다.	숙박시설의 위치는 쉽게 찾을 수 있고 한국의 대표적인 반지하 숙박시설입니다.	0.7428571428571428
  위반행위 조사 등을 거부·방해·기피한 자는 500만원 이하 과태료 부과 대상이다.	시민들 스스로 자발적인 예방 노력을 한 것은 아산 뿐만이 아니었다.	0.0
  회사가 보낸 메일은 이 지메일이 아니라 다른 지메일 계정으로 전달해줘.	사람들이 주로 네이버 메일을 쓰는 이유를 알려줘	0.06666666666666667

• Loss: CosineSimilarityLoss with these parameters:

  {
      "loss_fct": "torch.nn.modules.loss.MSELoss"
  }
  

Evaluation Dataset

klue/klue

• Dataset: klue/klue at 349481e
• Size: 519 evaluation samples
• Columns: sentence1, sentence2, and label
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	label
  type	string	string	float
  details	min: 7 tokens mean: 18.16 tokens max: 55 tokens	min: 7 tokens mean: 17.69 tokens max: 58 tokens	min: 0.0 mean: 0.5 max: 1.0

• Samples:

  sentence1	sentence2	label
  무엇보다도 호스트분들이 너무 친절하셨습니다.	무엇보다도, 호스트들은 매우 친절했습니다.	0.9714285714285713
  주요 관광지 모두 걸어서 이동가능합니다.	위치는 피렌체 중심가까지 걸어서 이동 가능합니다.	0.2857142857142858
  학생들의 균형 있는 영어능력을 향상시킬 수 있는 학교 수업을 유도하기 위해 2018학년도 수능부터 도입된 영어 영역 절대평가는 올해도 유지한다.	영어 영역의 경우 학생들이 한글 해석본을 암기하는 문제를 해소하기 위해 2016학년도부터 적용했던 EBS 연계 방식을 올해도 유지한다.	0.25714285714285723

• Loss: CosineSimilarityLoss with these parameters:

  {
      "loss_fct": "torch.nn.modules.loss.MSELoss"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• num_train_epochs: 30
• warmup_ratio: 0.1
• fp16: True

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
• 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: 30
• 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
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	loss	sts-dev_spearman_cosine
0	0	-	-	0.7123
0.0109	1	0.0255	-	-
0.5435	50	0.0225	0.0336	0.7961
1.0870	100	0.0159	0.0288	0.8299
1.6304	150	0.012	0.0258	0.8499
2.1739	200	0.0098	0.0238	0.8651
2.7174	250	0.0069	0.0233	0.8700
3.2609	300	0.0056	0.0241	0.8682
3.8043	350	0.0043	0.0231	0.8715
4.3478	400	0.0043	0.0261	0.8680
4.8913	450	0.0039	0.0239	0.8743
5.4348	500	0.0037	0.0247	0.8726
5.9783	550	0.0034	0.0231	0.8762
6.5217	600	0.003	0.0238	0.8746
7.0652	650	0.003	0.0246	0.8712
7.6087	700	0.0028	0.0240	0.8765

Framework Versions

• Python: 3.11.9
• Sentence Transformers: 3.0.1
• Transformers: 4.41.2
• PyTorch: 2.3.1
• Accelerate: 0.31.0
• Datasets: 2.19.2
• 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",
}