--- tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:47610 - loss:MultipleNegativesRankingLoss widget: - source_sentence: '[MENTION] Gustavus And Louise Pfeiffer Research Foundation [CITY] Bangor [COUNTRY] United States' sentences: - '[MENTION] Gustavus And Louise Pfeiffer Research Foundation [CITY] Bangor [COUNTRY] United States' - '[MENTION] Fifth Tianjin Central Hospital [CITY] Tianjin [COUNTRY] China' - '[MENTION] Purdue Research Foundation [ACRONYM] PRF [CITY] West Lafayette [COUNTRY] United States' - source_sentence: '[MENTION] ইন্টার-ইউরিভার্সিটি সেন্টার ফর অ্যাস্ট্রোনমি অ্যান্ড অ্যাস্ট্রোফিজিক্স [CITY] Pune [COUNTRY] India' sentences: - '[MENTION] National Centre for Radio Astrophysics [ACRONYM] NCRA TIFR [PARENT] Tata Institute of Fundamental Research [ACRONYM] TIFR [CITY] Pune [COUNTRY] India' - '[MENTION] Inter-University Centre for Astronomy and Astrophysics [ACRONYM] IUCAA [CITY] Pune [COUNTRY] India' - '[MENTION] Iskra Medical (Slovenia) [CITY] Radovljica [COUNTRY] Slovenia' - source_sentence: '[MENTION] Raytheon Technologies (Canada) [CITY] Calgary [COUNTRY] Canada' sentences: - '[MENTION] Raytheon Technologies (Canada) [ACRONYM] RCL [PARENT] RTX (United States) [CITY] Calgary [COUNTRY] Canada' - '[MENTION] Yunnan Open University [CITY] Kunming [COUNTRY] China' - '[MENTION] ATCO (Canada) [CITY] Calgary [COUNTRY] Canada' - source_sentence: '[MENTION] 유한양행 [CITY] Seoul' sentences: - '[MENTION] Instituto de Medicina Molecular João Lobo Antunes [ACRONYM] IMM [PARENT] University of Lisbon [CITY] Lisbon [COUNTRY] Portugal' - '[MENTION] Boehringer Ingelheim (South Korea) [PARENT] Boehringer Ingelheim (Germany) [CITY] Seoul [COUNTRY] South Korea' - '[MENTION] Yuhan (South Korea) [CITY] Seoul [COUNTRY] South Korea' - source_sentence: '[MENTION] Hyderabad Cleft Society [COUNTRY] India' sentences: - '[MENTION] Hyderabad Cleft Society [ACRONYM] HCS [CITY] Hyderabad [COUNTRY] India' - '[MENTION] Hyderabad Rheumatology Center [ACRONYM] HRC [CITY] Hyderabad [COUNTRY] India' - '[MENTION] National Institute of Technology Akita College [PARENT] National Institute of Technology [CITY] Akita [COUNTRY] Japan' pipeline_tag: sentence-similarity library_name: sentence-transformers metrics: - pearson_cosine - spearman_cosine model-index: - name: SentenceTransformer results: - task: type: semantic-similarity name: Semantic Similarity dataset: name: entity linking eval type: entity_linking_eval metrics: - type: pearson_cosine value: 0.7072780089709011 name: Pearson Cosine - type: spearman_cosine value: 0.6825742231480432 name: Spearman Cosine --- # SentenceTransformer This is a [sentence-transformers](https://www.SBERT.net) model trained. 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 - **Maximum Sequence Length:** 128 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': 128, 'do_lower_case': False}) with Transformer model: 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: ```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("SIRIS-Lab/affilgood-dense-retriever") # Run inference sentences = [ '[MENTION] Hyderabad Cleft Society [COUNTRY] India', '[MENTION] Hyderabad Cleft Society [ACRONYM] HCS [CITY] Hyderabad [COUNTRY] India', '[MENTION] Hyderabad Rheumatology Center [ACRONYM] HRC [CITY] Hyderabad [COUNTRY] India', ] 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: `entity_linking_eval` * Evaluated with [EmbeddingSimilarityEvaluator](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.EmbeddingSimilarityEvaluator) | Metric | Value | |:--------------------|:-----------| | pearson_cosine | 0.7073 | | **spearman_cosine** | **0.6826** | ## Training Details ### Training Dataset #### Unnamed Dataset * Size: 47,610 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 | | | | * Samples: | sentence_0 | sentence_1 | sentence_2 | |:--------------------------------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------| | [MENTION] The Prince Of Wales'S Institute Of Architecture [CITY] London [COUNTRY] United Kingdom | [MENTION] The Princes Foundation [CITY] London [COUNTRY] United Kingdom | [MENTION] Royal Institute of British Architects [ACRONYM] RIBA [CITY] London [COUNTRY] United Kingdom | | [MENTION] Development Finance & Public Policies [COUNTRY] Belgium | [MENTION] Development Finance and Public Policies [ACRONYM] DEFIPP [PARENT] University of Namur [CITY] Namur [COUNTRY] Belgium | [MENTION] Service Public Federal Finances [ACRONYM] SPF [CITY] Brussels [COUNTRY] Belgium | | [MENTION] EES [COUNTRY] United States | [MENTION] Emerald Education Systems [ACRONYM] EES [CITY] Pasadena [COUNTRY] United States | [MENTION] ESI Group (United States) [ACRONYM] ESI [PARENT] ESI Group (France) [ACRONYM] ESI [CITY] Farmington Hills [COUNTRY] United States | * Loss: [MultipleNegativesRankingLoss](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters: ```json { "scale": 20.0, "similarity_fct": "cos_sim" } ``` ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 16 - `per_device_eval_batch_size`: 16 - `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`: 16 - `per_device_eval_batch_size`: 16 - `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 - `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 - `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 - `prompts`: None - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: round_robin
### Training Logs | Epoch | Step | Training Loss | entity_linking_eval_spearman_cosine | |:------:|:----:|:-------------:|:-----------------------------------:| | 0.1680 | 500 | 0.3431 | - | | 0.3360 | 1000 | 0.252 | 0.4769 | | 0.5040 | 1500 | 0.291 | - | | 0.6720 | 2000 | 0.2445 | 0.6494 | | 0.8401 | 2500 | 0.2339 | - | | 1.0 | 2976 | - | 0.6694 | | 1.0081 | 3000 | 0.2256 | 0.6730 | | 1.1761 | 3500 | 0.16 | - | | 1.3441 | 4000 | 0.1428 | 0.6750 | | 1.5121 | 4500 | 0.1661 | - | | 1.6801 | 5000 | 0.139 | 0.6713 | | 1.8481 | 5500 | 0.1408 | - | | 2.0 | 5952 | - | 0.6768 | | 2.0161 | 6000 | 0.1409 | 0.6763 | | 2.1841 | 6500 | 0.0759 | - | | 2.3522 | 7000 | 0.0702 | 0.6820 | | 2.5202 | 7500 | 0.0716 | - | | 2.6882 | 8000 | 0.0777 | 0.6805 | | 2.8562 | 8500 | 0.0685 | - | | 3.0 | 8928 | - | 0.6826 | ### Framework Versions - Python: 3.10.12 - Sentence Transformers: 3.4.1 - Transformers: 4.41.2 - PyTorch: 2.2.0+cu121 - Accelerate: 1.2.1 - Datasets: 2.18.0 - Tokenizers: 0.19.1 ## 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", } ``` #### MultipleNegativesRankingLoss ```bibtex @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} } ```