m7n/discipline-tuned_specter_2_001

SentenceTransformer based on allenai/specter2_aug2023refresh_base

This is a sentence-transformers model finetuned from allenai/specter2_aug2023refresh_base. 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
• Base model: allenai/specter2_aug2023refresh_base
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 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': 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})
  (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("m7n/discipline-tuned_specter_2_001")
# Run inference
sentences = [
    'In this work, a novel design of an oscillating body-type Wave Energy Converter (WEC) is proposed with an efficiency of %. Innovative features of the new design include the integration of Mechanical Motion Rectifier (MMR), Motion Regulator (MR), Energy Storage Element (ESE), and Electric Generator (EG) with the operation controlled by a microcontroller, limit switches, and linear actuator. Lab experiments conducted with a prototype ensured a peak voltage of V and peak power of W. Experimental and MATLAB theoretical simulation results demonstrated the average power of - mW for a wave height of - mm. In comparison to the conventional operation, where the electric generator is directly driven by wave motion, the proposed design will ensure higher peak power and better energy utilization characteristics of the generated voltage waveform. Simulation results are presented with the incorporation of non-circular gears that ensure modification of the voltage waveform for better energy utilization.',
    'The concerns about the availability of freshwater to meet the demands of a growing population while sustaining a healthy natural environment are based on several factors: uncertainties as to the availability of supplies; the high costs of developing additional water supplies; the vulnerability of the resource and the problems of restoring and protecting valued surface and groundwater resources; the importance of reliable supplies of high-quality water for human and environmental health and economic development; and the shortcomings of our institutions for allocating scarce supplies in response to changing supply and demand conditions. Therefore water management authorities around the world are challenged with ensuring the quantity, quality, and allocation among the various uses of water are sustainable.This paper is based on the household survey of the irrigators using recycled water from the scheme to irrigate their crops in Virginia. The paper elicits their point of view on various issues related to wastewater usage and the rules-in-use governing wastewater management related to the scheme.DOI: Economic Journal of Development Issues Vol.00 & , pp.00-',
    "Objective: Use a systems-biology based modeling approach to identify and test mechanistic hypotheses that may explain the reductions in proteinuria and rate of GFR (Glomerular Filtration Rate) decline following the addition of aliskiren, a direct renin inhibitor (DRI), to Losartan, an angiotensin receptor blocker (ARB), as reported in the AVOID trial (00week study in diabetic nephropathy patients). Method: A RAAS-hypertension model was collaboratively developed by extending the well-established Guyton-Coleman model of blood pressure regulation to include ) a detailed representation of the renin angiotensin aldosterone (RAAS) pathway and ) a renal module that captures kidney damage caused by elevated glucose, renal angiotensin II, and blood pressure, leading to proteinuria and gradual decline in GFR. The model also includes a virtual patient (VP) population consisting of normotensives, hypertensives, and hypertensives with diabetic nephropathy. To parameterize the model, published and internal data on changes in RAAS biomarkers, blood pressure, GFR, and proteinuria following RAAS therapies were used. The AVOID study was simulated in the RAAS model using a cohort of diabetic nephropathy VPs selected to match baseline characteristics of patients in the trial. Results: The model was able to capture both the percent reduction in UACR (urinary albumin excretion rate) and differential rate of decline in GFR in the aliskiren arm compared to placebo observed in AVOID (Figure ). In this poster we present model-based hypothesis testing that points to multiple mechanistic factors that can contribute to added benefits of a DRI in combination with ARBs, including the degree of baseline intrarenal RAAS upregulation, the relative effectiveness of ARBs in inhibiting intrarenal versus systemic AT0 receptors, and variation in the population's sensitivity to RAAS therapies. Conclusions: We have developed a RAAS model that can be used to test hypotheses, to predict outcomes of future clinical trials, and to identify patient populations most likely to benefit from RAAS therapies.",
]
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

Triplet

• Datasets: specter_2_ and discipline-tuned_specter_2_001
• Evaluated with TripletEvaluator

Metric	specter_2_	discipline-tuned_specter_2_001
cosine_accuracy	0.9821	0.9821

Training Details

Training Dataset

Unnamed Dataset

• Size: 7,828 training samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 82 tokens mean: 232.84 tokens max: 512 tokens	min: 89 tokens mean: 236.44 tokens max: 512 tokens	min: 87 tokens mean: 238.23 tokens max: 512 tokens

• Samples:

  anchor	positive	negative
  Other	December Sir Thomas Wyatt and His Background. Patricia Thomson. Stanford: Stanford University Press, . xiv + pp. . Richard Harrier Richard Harrier Search for other works by this author on: This Site Google Modern Language Quarterly ( ) ( ): . Cite Icon Cite Share Icon Share Twitter Permissions Search Site Citation Richard Harrier; Sir Thomas Wyatt and His Background.. Modern Language Quarterly December ; ( ): . doi: Download citation file: Zotero Reference Manager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex toolbar search Search Dropdown Menu toolbar search search input Search input auto suggest filter your search Books & JournalsAll JournalsModern Language Quarterly Search Advanced Search The text of this article is only available as a PDF. Copyright ©️ by Duke University Press0000 Article PDF first page preview Close Modal Issue Section: Reviews You do not currently have access to this content.	Research Article
  Abstract In this paper, we aim to determine the importance of psychological pressure on performance in sequential tournaments. We make use of penalty shootouts in soccer and test the presence of firstmover advantage. Shootouts in soccer is a formidable framework because it is a natural experiment where teams compete in a sequence whose order is determined by the random outcome of a coin toss. Using data from all penalty shootouts in all major competitions since first implemented in , we fail to identify statistically significant differences in winning probabilities for teams kicking first or second in the shootout sequence.	Abstract We examine decisions made by village officials in a public goods game and their impact on villagers' attitudes toward a resettlement project in seven rural villages in China. The data comes from two sources: (i) a laboratory experiment with village officials, and (ii) a survey of villagers. Villages whose officials exhibited more cooperative behavior in the laboratory reported higher support levels from the villagers. We find that teams with higher average contribution levels in the PGG game relayed the resettlement news to their villagers earlier than other teams and also possess greater altruistic preferences.	Previous articleNext article No AccessRecentXiaoping Fang. China and the Cholera Pandemic: Restructuring Society under Mao. x + pp., notes, bibl., index. Pittsburgh: University of Pittsburgh Press, . (cloth); ISBN . Paper and e-book available.Ruth RogaskiRuth Rogaski Search for more articles by this author PDFPDF PLUSFull Text Add to favoritesDownload CitationTrack CitationsPermissionsReprints Share onFacebookTwitterLinkedInRedditEmail SectionsMoreDetailsFiguresReferencesCited by Isis Volume , Number 0June Publication of the History of Science Society Article DOIhttps://doi.org/ . Views: 00Total views on this site For permission to reuse, please contact [email protected]PDF download Crossref reports no articles citing this article.
  We used a computer data bank to evaluate consecutive patients admitted to a cardiac care unit with myocardial infarction. Stroke occurred in ( %) patients in the hospital; the anterior circulation was involved in % of strokes. Hospital mortality was % in patients with stroke and % in patients without stroke. Atrial arrhythmia was a significant (p ) risk factor for stroke, but peak creatine kinase and ventricular arrhythmia were not. Cardiac pump failure, apical or anterior-lateral myocardial infarction, and history of previous stroke were associated with an increased risk of stroke. Clinical and pathologic data suggested an embolic etiology for most strokes that complicate acute myocardial infarction.	Levetiracetam is a widely used antiseizure medication. Recent concerns have been raised regarding the potential prolongation of the QT interval by levetiracetam and increased risk of sudden cardiac death. This could have profound implications for patient safety and for prescribing practice. This study assessed the potential association of levetiracetam with cardiac outcomes related to QT interval prolongation. We compared outcomes of patients taking levetiracetam with those taking oxcarbazepine as a comparator medication that has not been associated with prolongation of the QT interval.	The density data for aqueous solutions of electrolytes have been analyzed, and partial molal volumes at infinite dilution have been calculated. The values are shown to be additive, and a set of volumes for individual ions has been prepared, based arbitrarily on a value of zero for the hydrogen ion. It is shown that for a given value of the charge the volumes vary linearly with the cube of the ionic crystal radii, and for a given radius vary with the first power of the charge. In the case of cations the equation obeyed is[Formula: see text]while for anions[Formula: see text]If the volume of the hydrogen ion is taken as ml. instead of zero the same equation is obeyed for both cations and anions, namely[Formula: see text]The empirical equations are discussed in terms of a simple model for ions in solution.

• Loss: TripletLoss with these parameters:

  {
      "distance_metric": "TripletDistanceMetric.COSINE",
      "triplet_margin": 0.3
  }
  

Evaluation Dataset

Unnamed Dataset

• Size: 391 evaluation samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 391 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 80 tokens mean: 231.7 tokens max: 512 tokens	min: 86 tokens mean: 234.38 tokens max: 512 tokens	min: 93 tokens mean: 233.73 tokens max: 512 tokens

• Samples:

  anchor	positive	negative
  Inorganic thermochromic materials exhibit a tunable color gamut and a wide chromatic temperature range, indicating their potential for intelligent adaptive applications in thermal warning, temperature indication, thermal regulation, and interactive light-to-thermal energy conversion. However, most metal-oxide-based thermochromic materials show weak chromaticity adaption with the change of temperature, which needs further understanding of the microscopic principle to clarify the potential route to improve the contrast and identifiability for fabricating better thermochromic materials. Using perovskite-structure (AMO0) alkaline earth metal stannate (Ba0xSrxSnO0, x ) as a model system, this paper reports for the first time the mechanism of the properties of thermally induced defect-enhanced charge transfer-type (CTT) thermochromic materials and the strategy for regulating their thermochromic properties by A-site cations. BaSnO0 exhibits continuously reversible thermochromic properties wit...	High-performance energy storage devices (HPEDs) play a critical role in the realization of clean energy and thus enable the overarching pursuit of nonpolluting, green technologies. Supercapacitors are one class of such lucrative HPEDs; however, a serious limiting factor of supercapacitor technology is its sub-par energy density. This report presents hitherto unchartered pathway of physical deformation, chemical dealloying, and microstructure engineering to produce ultrahigh-capacitance, energy-dense NiMn alloy electrodes. The activated electrode delivered an ultrahigh specific-capacitance of F/cm0 at A/cm0. The symmetric device showcased an excellent energy density of Wh/L and a remarkable cycle life of % retention after cycles. Transmission electron microscopy and atom probe tomography studies revealed the evolution of a unique hierarchical microstructure comprising fine Ni/NiMnO0 nanoligaments within MnO0-rich nanoflakes. Theoretical analysis using density functional theory showed se...	a peer-reviewed, open access online international journal which publishes original research papers. The journal welcomes submission from scholars and experts for possible publication from all over the world. The scope of the journal includes: Pharmaceutical research, chemistry and biochemistry of naturally occurring compounds, biological evaluation of crude extracts, ethnomedicine, traditional and complementary medicine, ethnopharmacology, biomedical research, Biotechnology, Evaluation of natural substances of land and sea and of plants, microbes and animals, pharmacognosy, bioavailability, clinical, pharmacological, toxicological studies and pharmacokinetics of phytochemicals, Isolation and characterization of compounds, structure elucidation, synthesis and experimental biosynthesis of natural Product as well as developments of methods in these areas are welcomed in the journal.
  Annals of the New York Academy of SciencesVolume , Issue p. - INTRODUCTION TO PARENTAL AGE AND CHARACTERISTICS OF THE OFFSPRING E. V. Cowdry, E. V. Cowdry Washington University, St. Louis, Mo.Search for more papers by this author E. V. Cowdry, E. V. Cowdry Washington University, St. Louis, Mo.Search for more papers by this author First published: January 0AboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL No abstract is available for this article.Citing Literature Volume00, Issue0Parental Age and Characteristics of OffspringJanuary 0000Pages - RelatedInformation	Annals of the New York Academy of SciencesVolume , Issue p. - Galactic Models with Moderate Stochasticity MARTIN SCHWARZSCHILD, MARTIN SCHWARZSCHILD Princeton University Observatory Princeton, New Jersey 00000Search for more papers by this author MARTIN SCHWARZSCHILD, MARTIN SCHWARZSCHILD Princeton University Observatory Princeton, New Jersey 00000Search for more papers by this author First published: May 0AboutPDF ToolsRequest permissionExport citationAdd to favoritesTrack citation ShareShare Give accessShare full text accessShare full-text accessPlease review our Terms and Conditions of Use and check box below to share full-text version of article.I have read and accept the Wiley Online Library Terms and Conditions of UseShareable LinkUse the link below to share a full-text version of this article with your friends and colleagues. Learn more.Copy URL Share a linkShare onFacebookTwitterLinked InRedditWechat Citing Literature Volume000, Issue0Chaotic Phenomena in AstrophysicsMay 0000Pa...	Nowadays,the main problems in financing of medium and small-sized enterprises are that there are some conflicts between family firm and equity financing,influenced by risk prevention and interest continuity;there is lack of trust between banks and enterprises -the threshold of indirect financing is too high through banks for the medium and small-sized enterprises,the banks are reluctant to loan due to fears of financing risk.These situations seriously limit development of medium and small-sized enterprises.Through establishing a game model,this paper analyzes a Bayesian Nash equilibrium on game between banks and enterprises on conditions of perfect information and imperfect information,and the study shows that information asymmetry is the main reason of hard financing.To come out of plight,the medium and small-sized enterprises should improve the mode of corporate governance and change family management, the government should play its macro-control role to enlarge financing channels an...
  Objective Carotid cavernous fistula (CCF) development after Pipeline Embolization Device (PED) treatment of cavernous carotid aneurysms (CCA) can be a challenging pathology to treat for the neurointerventionalist. Methods A database of all patients whose aneurysms were treated with the PED since its approval by the Food and Drug Administration in was retrospectively reviewed. Demographic information, aneurysm characteristics, treatment technique, antiplatelet regimen, and follow-up data were collected. A literature review of all papers that describe PED treatment of CCA was then completed. Results A total of patients with CCAs were identified ( women, men). The mean age was .0 years. The mean maximal aneurysm diameter was .0 mm (mean neck .0 mm). A single PED was deployed in patients, with two PEDs deployed in patients and three PEDs in patients. Adjunctive coiling was performed in patients. Mean follow-up duration based on final imaging (MR angiography or digital subtraction angiograp...	Cerebrospinal fluid-venous fistula is increasingly recognized as a cause of spontaneous intracranial hypotension.0 Transvenous embolization is emerging as an efficacious minimally invasive treatment.0- The procedure aims to embolize paraspinal and foraminal veins draining the fistula; however, complete embolization may be challenging as numerous small venous tributaries at the foraminal venous plexus, including dorsal muscular branches, may serve as additional routes of cerebrospinal fluid egress.0 To ensure curative embolization, we adopted a dual microcatheter pressure cooker technique, previously used for treatment of brain arteriovenous malformations.0 This allows improved control of embolic material reflux and greater chance of complete embolization of the site of the fistula and all potential venous tributaries. Video demonstrates this technique employed in a typical case using Onyx (Medtronic, Minnesota, USA) to embolize a cerebrospinal fluid-venous fistula at the left L0 neural...	In this report we compare the capabilities of two optical devicesa conventional opacity meter and a recently developed photoacoustic instrumentfor measurement of diesel particulate emissions. Emission measurements were performed using two vehicles (built in and and equipped with liter diesel engines) operated at a variety of steady-state conditions and also over the federal test procedure driving cycle. The light absorbed by diesel particles heats them thereby increasing the ambient gas pressure. The change in pressure, which can be related to the particle mass concentration, is measured in the photoacoustic instrument. This is about times more sensitive than the opacity meter so that it can be used to measure particulate emissions even in diluted exhaust. Further, its operation in the infrared region ensures that particle size variations do not affect its calibration against mass concentration. However, the observed optical data, both in the visible and in the infrared region, are dep...

• Loss: TripletLoss with these parameters:

  {
      "distance_metric": "TripletDistanceMetric.COSINE",
      "triplet_margin": 0.3
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 4
• per_device_eval_batch_size: 4
• learning_rate: 1e-05
• weight_decay: 0.01
• num_train_epochs: 2
• warmup_ratio: 0.1
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 4
• per_device_eval_batch_size: 4
• 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: 1e-05
• weight_decay: 0.01
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 2
• 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: 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: 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	specter_2__cosine_accuracy	discipline-tuned_specter_2_001_cosine_accuracy
0	0	-	-	0.9457	-
0.0511	100	0.1608	0.0894	0.9662	-
0.1022	200	0.0615	0.0412	0.9751	-
0.1533	300	0.0364	0.0328	0.9783	-
0.2044	400	0.0384	0.0289	0.9764	-
0.2555	500	0.0292	0.0245	0.9783	-
0.3066	600	0.0208	0.0262	0.9777	-
0.3577	700	0.0181	0.0276	0.9808	-
0.4088	800	0.023	0.0220	0.9834	-
0.4599	900	0.0254	0.0273	0.9764	-
0.5110	1000	0.021	0.0242	0.9796	-
0.5621	1100	0.0238	0.0245	0.9783	-
0.6132	1200	0.02	0.0258	0.9777	-
0.6643	1300	0.0284	0.0245	0.9725	-
0.7154	1400	0.0302	0.0198	0.9789	-
0.7665	1500	0.0229	0.0201	0.9821	-
0.8176	1600	0.0265	0.0196	0.9789	-
0.8687	1700	0.0185	0.0195	0.9777	-
0.9198	1800	0.0192	0.0212	0.9757	-
0.9709	1900	0.0342	0.0189	0.9840	-
1.0220	2000	0.0324	0.0190	0.9853	-
1.0731	2100	0.0174	0.0185	0.9815	-
1.1242	2200	0.0118	0.0188	0.9808	-
1.1753	2300	0.012	0.0202	0.9789	-
1.2264	2400	0.0061	0.0190	0.9802	-
1.2775	2500	0.0039	0.0188	0.9808	-
1.3286	2600	0.006	0.0197	0.9789	-
1.3797	2700	0.0047	0.0190	0.9815	-
1.4308	2800	0.0044	0.0187	0.9834	-
1.4819	2900	0.006	0.0189	0.9821	-
1.5330	3000	0.0042	0.0185	0.9821	-
1.5841	3100	0.005	0.0189	0.9815	-
1.6352	3200	0.004	0.0195	0.9808	-
1.6863	3300	0.0034	0.0195	0.9828	-
1.7374	3400	0.0061	0.0196	0.9789	-
1.7885	3500	0.0043	0.0192	0.9815	-
1.8396	3600	0.0062	0.0187	0.9802	-
1.8906	3700	0.0036	0.0187	0.9815	-
1.9417	3800	0.0077	0.0188	0.9821	-
1.9928	3900	0.0139	0.0188	0.9821	-
2.0	3914	-	-	-	0.9821

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.3.1
• Transformers: 4.48.0.dev0
• PyTorch: 2.5.1+cu121
• Accelerate: 1.2.1
• Datasets: 3.2.0
• Tokenizers: 0.21.0

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}
}