metadata

library_name: sentence-transformers
metrics:
  - pearson_cosine
  - spearman_cosine
  - pearson_manhattan
  - spearman_manhattan
  - pearson_euclidean
  - spearman_euclidean
  - pearson_dot
  - spearman_dot
  - pearson_max
  - spearman_max
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:10501
  - loss:CosineSimilarityLoss
widget:
  - source_sentence: 추운날이니 외출은 자제해주시기 바랍니다.
    sentences:
      - 추운날인데 외출하지마
      - 소·돼지에 대해서만 실시하던 축산물이력제가 1월 1일부터 닭·오리·계란까지 확대·시행된다.
      - 광고메일함 비중이 에어비앤비가 더 높니 트립닷컴이 더 많니?
  - source_sentence: 샤워기도 수압이 너무 약해서 불편해요.
    sentences:
      - 숙소 내부가 넓고 호스트도 1층에 있어 불편사항에 대한 피드백을 즉시 받으실 수 있습니다.
      - >-
        그외에 물놀이를 하기위한 준비물들 파라솔 비치의자 어린이비치의자 아이스박스 핸드케리어 비치타월 모레놀이도구 등등 필요한 모든것이
        완벽했습니다.
      - 샤워는 수압이 너무 약해서 불편해요.
  - source_sentence: 조용한 분위기의 방을 구하시면 이 곳이 최고입니다!
    sentences:
      - 시험을 이번달에 본다고 했니 다음달에 본다고 했니?
      - 조용한 방을 찾는다면, 이곳이 최고예요!
      - 어른들과 만나는 자리에는 어른들보다 늦게 도착하지 말고 일찍 나가 있어라.
  - source_sentence: 발코니쪽 창문은 3개중에 한개만 열수있습니다.
    sentences:
      - 많은 장비를 구매할 필요 없이 즐길 수 있습니다.
      - 우리는 그 숙소에서 호바트의 최상의 상태를 유지할 수 있었습니다.
      - 직장가입자의 급여명세서, 지역가입자의 건강보험 급여통지서를 확인하실 수 있습니다.
  - source_sentence: 국민 추천으로 ‘금융규제 유연화로 선제적 금융권 지원역량 강화’도 우수 사례로 언급됐다.
    sentences:
      - 국민의 권고에 따라 '유연한 금융규제 등을 통해 선제적으로 금융분야 지원능력 강화'도 좋은 사례로 꼽혔습니다.
      - 사진으로 보이는거 보다 숙소는 넓었고요
      - 저는 다음에 대만을 간다면 무조건 재방문 할 예정입니다!
model-index:
  - name: SentenceTransformer
    results:
      - task:
          type: semantic-similarity
          name: Semantic Similarity
        dataset:
          name: sts dev
          type: sts-dev
        metrics:
          - type: pearson_cosine
            value: 0.9626619602187976
            name: Pearson Cosine
          - type: spearman_cosine
            value: 0.9247880695962829
            name: Spearman Cosine
          - type: pearson_manhattan
            value: 0.9555167285690431
            name: Pearson Manhattan
          - type: spearman_manhattan
            value: 0.923408354022865
            name: Spearman Manhattan
          - type: pearson_euclidean
            value: 0.9556439523907834
            name: Pearson Euclidean
          - type: spearman_euclidean
            value: 0.9235806565450854
            name: Spearman Euclidean
          - type: pearson_dot
            value: 0.957361957340705
            name: Pearson Dot
          - type: spearman_dot
            value: 0.9130155209197447
            name: Spearman Dot
          - type: pearson_max
            value: 0.9626619602187976
            name: Pearson Max
          - type: spearman_max
            value: 0.9247880695962829
            name: Spearman Max

SentenceTransformer

This is a sentence-transformers model trained. 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: 256 tokens
Output Dimensionality: 768 tokens
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': 256, 'do_lower_case': True}) with Transformer model: RobertaModel 
  (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 = [
    '국민 추천으로 ‘금융규제 유연화로 선제적 금융권 지원역량 강화’도 우수 사례로 언급됐다.',
    "국민의 권고에 따라 '유연한 금융규제 등을 통해 선제적으로 금융분야 지원능력 강화'도 좋은 사례로 꼽혔습니다.",
    '사진으로 보이는거 보다 숙소는 넓었고요',
]
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	0.9627
spearman_cosine	0.9248
pearson_manhattan	0.9555
spearman_manhattan	0.9234
pearson_euclidean	0.9556
spearman_euclidean	0.9236
pearson_dot	0.9574
spearman_dot	0.913
pearson_max	0.9627
spearman_max	0.9248

Training Details

Training Dataset

Unnamed Dataset

Size: 10,501 training samples
Columns: sentence_0, sentence_1, and label
Approximate statistics based on the first 1000 samples:
sentence_0 sentence_1 label
type string string float
details
min: 5 tokens
mean: 20.16 tokens
max: 58 tokens

min: 6 tokens
mean: 19.75 tokens
max: 58 tokens

min: 0.0
mean: 0.44
max: 1.0

	sentence_0	sentence_1	label
type	string	string	float
details	min: 5 tokens mean: 20.16 tokens max: 58 tokens	min: 6 tokens mean: 19.75 tokens max: 58 tokens	min: 0.0 mean: 0.44 max: 1.0

Samples:

sentence_0	sentence_1	label
`단점을 꼽자면 엘베가 없다는 점 정도?`	`굳이 단점을 꼽자면 늦은 밤에는 역 근처가 살짝 무섭다는 거?`	`0.2`
`더울 때는 청량음료 말고 물 많이 마셔.`	`추울 때 손과 발은 내놓지 말자.`	`0.0`
`위치, 시설, 호스팅 모두 만족했습니다.`	`위치, 시설, 호스팅 모두 만족스러웠습니다.`	`1.0`

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: 32
per_device_eval_batch_size: 32
num_train_epochs: 7
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: 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: 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: 7
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
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: batch_sampler
multi_dataset_batch_sampler: round_robin

Training Logs

Epoch	Step	Training Loss	sts-dev_spearman_max
1.0	329	-	0.9218
1.5198	500	0.0096	-
2.0	658	-	0.9218
3.0	987	-	0.9215
3.0395	1000	0.0064	0.9218
4.0	1316	-	0.9231
4.5593	1500	0.0055	-
5.0	1645	-	0.9231
6.0	1974	-	0.9235
6.0790	2000	0.0045	0.9226
7.0	2303	-	0.9248

Framework Versions

Python: 3.10.12
Sentence Transformers: 3.1.1
Transformers: 4.44.2
PyTorch: 2.4.1+cu121
Accelerate: 0.34.2
Datasets: 3.0.1
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",
}