SentenceTransformer based on juyoungml/bge-m3-ko-v1.1

This is a sentence-transformers model finetuned from juyoungml/bge-m3-ko-v1.1. 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: juyoungml/bge-m3-ko-v1.1
Maximum Sequence Length: 8192 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': 8192, 'do_lower_case': False, 'architecture': 'XLMRobertaModel'})
  (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, '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 = [
    '제주 항일운동가 손실 소식에 대해 알고 싶습니다.',
    '다. 항일ㆍ노동운동의 거목이 손실되자 조선인ㆍ일본인의 노동계가 합동으로 시신을 유리관(棺)에 안치, 오사카시 노동장(勞動葬)으로 성대히 엄수... (내용이 길어 일부 생략됨)',
    '머리가 되는 것이 아니다. 누구가 먼저 키를 잡는 수완과 자질을 발휘해 선점하느냐에 따라 도사공의 지위를 차지하게 된다는 말이다.',
]
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, -0.9309, -0.9304],
#         [-0.9309,  1.0000,  0.9989],
#         [-0.9304,  0.9989,  1.0000]])

Evaluation

Metrics

Triplet

Dataset: val
Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.4135

Training Details

Training Dataset

Unnamed Dataset

Size: 6,238 training samples
Columns: sentence_0, sentence_1, and sentence_2

Approximate statistics based on the first 1000 samples:

	sentence_0	sentence_1	sentence_2
type	string	string	string
details	min: 8 tokens mean: 20.85 tokens max: 91 tokens	min: 8 tokens mean: 216.35 tokens max: 2850 tokens	min: 6 tokens mean: 193.04 tokens max: 2850 tokens

Samples:

sentence_0	sentence_1	sentence_2
`할머니, 할아버지는 어떤 모습일까요?`	`[삼 서끈 할망은 놀레 불르곡, 노 서끈 하르방은 운다.] 삼 섞은 할머니는 노래 부르고, 노 섞은 할아버지는 운다. 여자는 인내력과 서둘지 않은 차분한 일 처리 능력이 남자에 비해서 앞선다는 말이다.`	`정부에서는 고인의 공훈을 기리어 1990년 광복절에 건국훈장 애족장을 추서하였다.`
`제주도에서는 밥 먹는 일이 일과 어떤 연관이 있나요?`	하고 다음 일을 할 수 있다. 어느 한 쪽이 늦어지면 그만큼 다음 일 착수가 늦어질 뿐만 아리나 일 처리도 미뤄지므로, 밥 먹는 일 자체가 곧 일을 차리는 것이 된다는 말이다. 의미: 밥 먹음도 일 차림이다. 밥을 먹는 일은 평상시 정례화 된 사실이지만, 먹는 쪽이나 그 식사를 마련하는 쪽이 다 정해진 일과의 한 과정을 수행하고 있는 것이다. 그러니 먹는 사람은 먹어야 다음 일을 시작할 수 있고, 식사를 차려놓는 사람은 제 때에 먹어 버려야 뒤처리를 하고 다음 일을 할 수 있다. 어느 한 쪽이 늦어지면 그만큼 다음 일 착수가 늦어질 뿐만 아리나 일 처리도 미뤄지므로, 밥 먹는 일 자체가 곧 일을 차리는 것이 된다는 말이다.	나하고 둘이 둥실 사랑하자.`
`행원남당본풀이 이야기 들어보고 싶습니다.`	`[행원남당본풀이] (myth) [신화] MYTH2.002 - 핵심어: 본풀이,행원남당본풀이`	`못하고 몸만 상하고 마는 현상을 지적하고 있다`

Loss: TripletLoss with these parameters:

{
    "distance_metric": "TripletDistanceMetric.EUCLIDEAN",
    "triplet_margin": 5
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 1
per_device_eval_batch_size: 1
fp16: True
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: 1
per_device_eval_batch_size: 1
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: 3
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
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}
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
project: huggingface
trackio_space_id: trackio
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: no
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: True
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: round_robin
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	val_cosine_accuracy
0.0802	500	4.8733	0.4135

Framework Versions

Python: 3.12.12
Sentence Transformers: 5.2.0
Transformers: 4.57.3
PyTorch: 2.9.0+cu126
Accelerate: 1.12.0
Datasets: 4.0.0
Tokenizers: 0.22.1

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",
}

TripletLoss

@misc{hermans2017defense,
    title={In Defense of the Triplet Loss for Person Re-Identification},
    author={Alexander Hermans and Lucas Beyer and Bastian Leibe},
    year={2017},
    eprint={1703.07737},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}

Downloads last month: -

Safetensors

Model size

0.6B params

Tensor type

F32

Model tree for JHyeok5/aewol-embedding-v1

Base model

BAAI/bge-m3

Finetuned

juyoungml/bge-m3-ko-v1.1

Finetuned

(1)

this model

Papers for JHyeok5/aewol-embedding-v1

Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks

Paper • 1908.10084 • Published Aug 27, 2019 • 12

In Defense of the Triplet Loss for Person Re-Identification

Paper • 1703.07737 • Published Mar 22, 2017

Evaluation results

Cosine Accuracy on val
self-reported

0.414