--- tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:5749 - loss:CosineSimilarityLoss base_model: CocoRoF/mobert_retry_SimCSE_test widget: - source_sentence: 우리는 움직이는 동행 우주 정지 좌표계에 비례하여 이동하고 있습니다 ... 약 371km / s에서 별자리 leo 쪽으로. " sentences: - 두 마리의 독수리가 가지에 앉는다. - 다른 물체와는 관련이 없는 '정지'는 없다. - 소녀는 버스의 열린 문 앞에 서 있다. - source_sentence: 숲에는 개들이 있다. sentences: - 양을 보는 아이들. - 여왕의 배우자를 "왕"이라고 부르지 않는 것은 아주 좋은 이유가 있다. 왜냐하면 그들은 왕이 아니기 때문이다. - 개들은 숲속에 혼자 있다. - source_sentence: '첫째, 두 가지 다른 종류의 대시가 있다는 것을 알아야 합니다 : en 대시와 em 대시.' sentences: - 그들은 그 물건들을 집 주변에 두고 가거나 집의 정리를 해칠 의도가 없다. - 세미콜론은 혼자 있을 수 있는 문장에 참여하는데 사용되지만, 그들의 관계를 강조하기 위해 결합됩니다. - 그의 남동생이 지켜보는 동안 집 앞에서 트럼펫을 연주하는 금발의 아이. - source_sentence: 한 여성이 생선 껍질을 벗기고 있다. sentences: - 한 남자가 수영장으로 뛰어들었다. - 한 여성이 프라이팬에 노란 혼합물을 부어 넣고 있다. - 두 마리의 갈색 개가 눈 속에서 서로 놀고 있다. - source_sentence: 버스가 바쁜 길을 따라 운전한다. sentences: - 우리와 같은 태양계가 은하계 밖에서 존재할 수도 있을 것입니다. - 그 여자는 데이트하러 가는 중이다. - 녹색 버스가 도로를 따라 내려간다. pipeline_tag: sentence-similarity library_name: sentence-transformers metrics: - pearson_cosine - spearman_cosine - pearson_euclidean - spearman_euclidean - pearson_manhattan - spearman_manhattan - pearson_dot - spearman_dot - pearson_max - spearman_max model-index: - name: SentenceTransformer based on CocoRoF/mobert_retry_SimCSE_test results: - task: type: semantic-similarity name: Semantic Similarity dataset: name: sts dev type: sts_dev metrics: - type: pearson_cosine value: 0.7885728442437165 name: Pearson Cosine - type: spearman_cosine value: 0.7890106880187878 name: Spearman Cosine - type: pearson_euclidean value: 0.7209624590910948 name: Pearson Euclidean - type: spearman_euclidean value: 0.7132906703480484 name: Spearman Euclidean - type: pearson_manhattan value: 0.7228003273015342 name: Pearson Manhattan - type: spearman_manhattan value: 0.7161151111265872 name: Spearman Manhattan - type: pearson_dot value: 0.7119673656141701 name: Pearson Dot - type: spearman_dot value: 0.7059066541365785 name: Spearman Dot - type: pearson_max value: 0.7885728442437165 name: Pearson Max - type: spearman_max value: 0.7890106880187878 name: Spearman Max --- # SentenceTransformer based on CocoRoF/mobert_retry_SimCSE_test This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [CocoRoF/mobert_retry_SimCSE_test](https://huggingface.co/CocoRoF/mobert_retry_SimCSE_test). 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:** [CocoRoF/mobert_retry_SimCSE_test](https://huggingface.co/CocoRoF/mobert_retry_SimCSE_test) - **Maximum Sequence Length:** 2048 tokens - **Output Dimensionality:** 1024 dimensions - **Similarity Function:** Cosine Similarity ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers) - **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers) ### Full Model Architecture ``` SentenceTransformer( (0): Transformer({'max_seq_length': 2048, 'do_lower_case': False}) with Transformer model: ModernBertModel (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}) (2): Dense({'in_features': 768, 'out_features': 1024, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'}) ) ``` ## Usage ### Direct Usage (Sentence Transformers) First install the Sentence Transformers library: ```bash pip install -U sentence-transformers ``` Then you can load this model and run inference. ```python from sentence_transformers import SentenceTransformer # Download from the 🤗 Hub model = SentenceTransformer("CocoRoF/ModernBERT-SimCSE-multitask_v03-retry") # Run inference sentences = [ '버스가 바쁜 길을 따라 운전한다.', '녹색 버스가 도로를 따라 내려간다.', '그 여자는 데이트하러 가는 중이다.', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 1024] # 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](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator) | Metric | Value | |:-------------------|:----------| | pearson_cosine | 0.7886 | | spearman_cosine | 0.789 | | pearson_euclidean | 0.721 | | spearman_euclidean | 0.7133 | | pearson_manhattan | 0.7228 | | spearman_manhattan | 0.7161 | | pearson_dot | 0.712 | | spearman_dot | 0.7059 | | pearson_max | 0.7886 | | **spearman_max** | **0.789** | ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 5,749 training samples * Columns: sentence1, sentence2, and score * Approximate statistics based on the first 1000 samples: | | sentence1 | sentence2 | score | |:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------| | type | string | string | float | | details | | | | * Samples: | sentence1 | sentence2 | score | |:------------------------------------|:------------------------------------------|:------------------| | 비행기가 이륙하고 있다. | 비행기가 이륙하고 있다. | 1.0 | | 한 남자가 큰 플루트를 연주하고 있다. | 남자가 플루트를 연주하고 있다. | 0.76 | | 한 남자가 피자에 치즈를 뿌려놓고 있다. | 한 남자가 구운 피자에 치즈 조각을 뿌려놓고 있다. | 0.76 | * Loss: [CosineSimilarityLoss](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosinesimilarityloss) with these parameters: ```json { "loss_fct": "torch.nn.modules.loss.MSELoss" } ``` ### Evaluation Dataset #### Unnamed Dataset * Size: 1,500 evaluation samples * Columns: sentence1, sentence2, and score * Approximate statistics based on the first 1000 samples: | | sentence1 | sentence2 | score | |:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------| | type | string | string | float | | details | | | | * Samples: | sentence1 | sentence2 | score | |:-------------------------------------|:------------------------------------|:------------------| | 안전모를 가진 한 남자가 춤을 추고 있다. | 안전모를 쓴 한 남자가 춤을 추고 있다. | 1.0 | | 어린아이가 말을 타고 있다. | 아이가 말을 타고 있다. | 0.95 | | 한 남자가 뱀에게 쥐를 먹이고 있다. | 남자가 뱀에게 쥐를 먹이고 있다. | 1.0 | * Loss: [CosineSimilarityLoss](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#cosinesimilarityloss) with these parameters: ```json { "loss_fct": "torch.nn.modules.loss.MSELoss" } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `overwrite_output_dir`: True - `eval_strategy`: steps - `per_device_train_batch_size`: 1 - `per_device_eval_batch_size`: 1 - `gradient_accumulation_steps`: 16 - `learning_rate`: 8e-05 - `num_train_epochs`: 10.0 - `warmup_ratio`: 0.2 - `push_to_hub`: True - `hub_model_id`: CocoRoF/ModernBERT-SimCSE-multitask_v03-retry - `hub_strategy`: checkpoint - `batch_sampler`: no_duplicates #### All Hyperparameters
Click to expand - `overwrite_output_dir`: True - `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`: 16 - `eval_accumulation_steps`: None - `torch_empty_cache_steps`: None - `learning_rate`: 8e-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`: 10.0 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.2 - `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`: True - `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`: True - `resume_from_checkpoint`: None - `hub_model_id`: CocoRoF/ModernBERT-SimCSE-multitask_v03-retry - `hub_strategy`: checkpoint - `hub_private_repo`: None - `hub_always_push`: False - `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 - `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 - `use_liger_kernel`: False - `eval_use_gather_object`: False - `average_tokens_across_devices`: False - `prompts`: None - `batch_sampler`: no_duplicates - `multi_dataset_batch_sampler`: proportional
### Training Logs | Epoch | Step | Training Loss | Validation Loss | sts_dev_spearman_max | |:------:|:----:|:-------------:|:---------------:|:--------------------:| | 0.1114 | 5 | - | 0.0377 | 0.7471 | | 0.2228 | 10 | 0.6923 | 0.0377 | 0.7471 | | 0.3343 | 15 | - | 0.0376 | 0.7473 | | 0.4457 | 20 | 0.6832 | 0.0376 | 0.7475 | | 0.5571 | 25 | - | 0.0375 | 0.7479 | | 0.6685 | 30 | 0.6787 | 0.0375 | 0.7484 | | 0.7799 | 35 | - | 0.0374 | 0.7488 | | 0.8914 | 40 | 0.6154 | 0.0373 | 0.7494 | | 1.0223 | 45 | - | 0.0372 | 0.7500 | | 1.1337 | 50 | 0.6231 | 0.0371 | 0.7506 | | 1.2451 | 55 | - | 0.0370 | 0.7512 | | 1.3565 | 60 | 0.6562 | 0.0369 | 0.7519 | | 1.4680 | 65 | - | 0.0368 | 0.7526 | | 1.5794 | 70 | 0.6578 | 0.0366 | 0.7534 | | 1.6908 | 75 | - | 0.0365 | 0.7541 | | 1.8022 | 80 | 0.6669 | 0.0364 | 0.7549 | | 1.9136 | 85 | - | 0.0363 | 0.7559 | | 2.0446 | 90 | 0.6428 | 0.0361 | 0.7568 | | 2.1560 | 95 | - | 0.0360 | 0.7577 | | 2.2674 | 100 | 0.5854 | 0.0358 | 0.7586 | | 2.3788 | 105 | - | 0.0357 | 0.7597 | | 2.4903 | 110 | 0.6027 | 0.0356 | 0.7607 | | 2.6017 | 115 | - | 0.0354 | 0.7618 | | 2.7131 | 120 | 0.6375 | 0.0353 | 0.7627 | | 2.8245 | 125 | - | 0.0351 | 0.7635 | | 2.9359 | 130 | 0.6204 | 0.0350 | 0.7643 | | 3.0669 | 135 | - | 0.0348 | 0.7653 | | 3.1783 | 140 | 0.6077 | 0.0347 | 0.7663 | | 3.2897 | 145 | - | 0.0346 | 0.7672 | | 3.4011 | 150 | 0.5772 | 0.0344 | 0.7681 | | 3.5125 | 155 | - | 0.0343 | 0.7690 | | 3.6240 | 160 | 0.5793 | 0.0341 | 0.7698 | | 3.7354 | 165 | - | 0.0340 | 0.7705 | | 3.8468 | 170 | 0.5807 | 0.0338 | 0.7712 | | 3.9582 | 175 | - | 0.0337 | 0.7721 | | 4.0891 | 180 | 0.5576 | 0.0336 | 0.7729 | | 4.2006 | 185 | - | 0.0334 | 0.7734 | | 4.3120 | 190 | 0.5244 | 0.0333 | 0.7740 | | 4.4234 | 195 | - | 0.0332 | 0.7748 | | 4.5348 | 200 | 0.539 | 0.0331 | 0.7754 | | 4.6462 | 205 | - | 0.0330 | 0.7760 | | 4.7577 | 210 | 0.5517 | 0.0329 | 0.7765 | | 4.8691 | 215 | - | 0.0328 | 0.7769 | | 4.9805 | 220 | 0.5265 | 0.0327 | 0.7776 | | 5.1114 | 225 | - | 0.0326 | 0.7780 | | 5.2228 | 230 | 0.5285 | 0.0325 | 0.7783 | | 5.3343 | 235 | - | 0.0324 | 0.7789 | | 5.4457 | 240 | 0.4697 | 0.0323 | 0.7793 | | 5.5571 | 245 | - | 0.0323 | 0.7798 | | 5.6685 | 250 | 0.4913 | 0.0322 | 0.7804 | | 5.7799 | 255 | - | 0.0321 | 0.7809 | | 5.8914 | 260 | 0.5253 | 0.0320 | 0.7813 | | 6.0223 | 265 | - | 0.0320 | 0.7817 | | 6.1337 | 270 | 0.4924 | 0.0319 | 0.7819 | | 6.2451 | 275 | - | 0.0318 | 0.7820 | | 6.3565 | 280 | 0.4844 | 0.0317 | 0.7822 | | 6.4680 | 285 | - | 0.0317 | 0.7825 | | 6.5794 | 290 | 0.442 | 0.0316 | 0.7827 | | 6.6908 | 295 | - | 0.0315 | 0.7830 | | 6.8022 | 300 | 0.4665 | 0.0314 | 0.7834 | | 6.9136 | 305 | - | 0.0314 | 0.7839 | | 7.0446 | 310 | 0.4672 | 0.0314 | 0.7843 | | 7.1560 | 315 | - | 0.0314 | 0.7851 | | 7.2674 | 320 | 0.4131 | 0.0314 | 0.7850 | | 7.3788 | 325 | - | 0.0313 | 0.7849 | | 7.4903 | 330 | 0.4221 | 0.0312 | 0.7848 | | 7.6017 | 335 | - | 0.0311 | 0.7854 | | 7.7131 | 340 | 0.4268 | 0.0310 | 0.7857 | | 7.8245 | 345 | - | 0.0309 | 0.7861 | | 7.9359 | 350 | 0.4316 | 0.0309 | 0.7866 | | 8.0669 | 355 | - | 0.0309 | 0.7872 | | 8.1783 | 360 | 0.4277 | 0.0309 | 0.7873 | | 8.2897 | 365 | - | 0.0308 | 0.7870 | | 8.4011 | 370 | 0.3925 | 0.0308 | 0.7868 | | 8.5125 | 375 | - | 0.0308 | 0.7866 | | 8.6240 | 380 | 0.4049 | 0.0308 | 0.7869 | | 8.7354 | 385 | - | 0.0308 | 0.7875 | | 8.8468 | 390 | 0.3742 | 0.0308 | 0.7883 | | 8.9582 | 395 | - | 0.0307 | 0.7885 | | 9.0891 | 400 | 0.3498 | 0.0307 | 0.7886 | | 9.2006 | 405 | - | 0.0307 | 0.7881 | | 9.3120 | 410 | 0.3569 | 0.0307 | 0.7878 | | 9.4234 | 415 | - | 0.0307 | 0.7876 | | 9.5348 | 420 | 0.3312 | 0.0306 | 0.7877 | | 9.6462 | 425 | - | 0.0305 | 0.7881 | | 9.7577 | 430 | 0.3848 | 0.0304 | 0.7885 | | 9.8691 | 435 | - | 0.0304 | 0.7889 | | 9.9805 | 440 | 0.332 | 0.0305 | 0.7890 | ### Framework Versions - Python: 3.11.10 - Sentence Transformers: 3.4.1 - Transformers: 4.48.3 - PyTorch: 2.5.1+cu124 - Accelerate: 1.3.0 - Datasets: 3.3.0 - Tokenizers: 0.21.0 ## Citation ### BibTeX #### Sentence Transformers ```bibtex @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", } ```