Abeshith/research-embedding-finetuned

SentenceTransformer based on sentence-transformers/all-mpnet-base-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-mpnet-base-v2 on the research-embedding-dataset dataset. 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: sentence-transformers/all-mpnet-base-v2
• Maximum Sequence Length: 384 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • research-embedding-dataset

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': 384, 'do_lower_case': False, 'architecture': 'MPNetModel'})
  (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("Abeshith/research-embedding-finetuned")
# Run inference
queries = [
    "01155e5dc07f948d1b2fb1ad92ec50ae",
]
documents = [
    'mteb/scifact',
    'Human embryonic stem cells give rise to cell types from the outer embryonic germ layer, but not the other two layers.',
    'Evidence of a pluripotent human embryonic stem cell line derived from a cloned blastocyst.. Somatic cell nuclear transfer (SCNT) technology has recently been used to generate animals with a common genetic composition. In this study, we report the derivation of a pluripotent embryonic stem (ES) cell line (SCNT-hES-1) from a cloned human blastocyst. The SCNT-hES-1 cells displayed typical ES cell morphology and cell surface markers and were capable of differentiating into embryoid bodies in vitro and of forming teratomas in vivo containing cell derivatives from all three embryonic germ layers in severe combined immunodeficient mice. After continuous proliferation for more than 70 passages, SCNT-hES-1 cells maintained normal karyotypes and were genetically identical to the somatic nuclear donor cells. Although we cannot completely exclude the possibility that the cells had a parthenogenetic origin, imprinting analyses support a SCNT origin of the derived human ES cells',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 768] [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)
# tensor([[ 0.8614, -0.0783, -0.0902]])

Training Details

Training Dataset

research-embedding-dataset

• Dataset: research-embedding-dataset at 4dc4ce9
• Size: 1,210 training samples
• Columns: id, source_dataset, query, positive, negatives, query_id, positive_id, negative_ids, task_type, and domain
• Approximate statistics based on the first 1000 samples:

  	id	source_dataset	query	positive	negatives	query_id	positive_id	negative_ids	task_type	domain
  type	string	string	string	string	list	string	string	list	string	string
  details	min: 20 tokens mean: 25.15 tokens max: 31 tokens	min: 8 tokens mean: 8.51 tokens max: 10 tokens	min: 7 tokens mean: 20.84 tokens max: 79 tokens	min: 5 tokens mean: 285.48 tokens max: 384 tokens	size: 4 elements	min: 3 tokens mean: 3.5 tokens max: 4 tokens	min: 4 tokens mean: 6.6 tokens max: 10 tokens	size: 4 elements	min: 7 tokens mean: 7.0 tokens max: 7 tokens	min: 3 tokens mean: 3.0 tokens max: 3 tokens

• Samples:

  id	source_dataset	query	positive	negatives	query_id	positive_id	negative_ids	task_type	domain
  db9464f292a9ee27df30636954660d2c	mteb/scifact	Primary cervical cytology screening with HPV detection has higher longitudinal sensitivity to detect severe cervical intraepithelial neoplasia than conventional cytology.	Rate of cervical cancer, severe intraepithelial neoplasia, and adenocarcinoma in situ in primary HPV DNA screening with cytology triage: randomised study within organised screening programme.. OBJECTIVE To assess the performance and impact of primary human papillomavirus (HPV) DNA screening with cytology triage compared with conventional cytology on cervical cancer and severe pre-cancerous lesions. DESIGN Randomised trial. SETTING Population based screening programme for cervical cancer in southern Finland in 2003-5. PARTICIPANTS 58 076 women, aged 30-60, invited to the routine population based screening programme for cervical cancer. INTERVENTIONS Primary HPV DNA test (hybrid capture II) with cytology triage if the result was positive or conventional cytological screening (reference). MAIN OUTCOME MEASURES Rate of cervical cancer, cervical intraepithelial neoplasia (CIN) grade III, and adenocarcinoma in situ (as a composite outcome referred to as CIN III+) during 2003-7 through record...	['Efficacy of HPV DNA testing with cytology triage and/or repeat HPV DNA testing in primary cervical cancer screening.. BACKGROUND Primary cervical screening with both human papillomavirus (HPV) DNA testing and cytological examination of cervical cells with a Pap test (cytology) has been evaluated in randomized clinical trials. Because the vast majority of women with positive cytology are also HPV DNA positive, screening strategies that use HPV DNA testing as the primary screening test may be more effective. METHODS We used the database from the intervention arm (n = 6,257 women) of a population-based randomized trial of double screening with cytology and HPV DNA testing to evaluate the efficacy of 11 possible cervical screening strategies that are based on HPV DNA testing alone, cytology alone, and HPV DNA testing combined with cytology among women aged 32-38 years. The main outcome measures were sensitivity for detection of cervical intraepithelial neoplasia grade 3 or worse (CIN3+) within 6 months of enrollment or at colposcopy for women with a persistent type-specific HPV infection and the number of screening tests and positive predictive value (PPV) for each screening strategy. All statistical tests were two-sided. RESULTS Compared with screening by cytology alone, double testing with cytology and for type-specific HPV persistence resulted in a 35% (95% confidence interval [CI] = 15% to 60%) increase in sensitivity to detect CIN3+, without a statistically significant reduction in the PPV (relative PPV = 0.76, 95% CI = 0.52 to 1.10), but with more than twice as many screening tests needed. Several strategies that incorporated screening for high-risk HPV subtypes were explored, but they resulted in reduced PPV compared with cytology. Compared with cytology, primary', 'Evaluation of human papillomavirus testing in primary screening for cervical abnormalities: comparison of sensitivity, specificity, and frequency of referral.. CONTEXT Human papillomavirus (HPV) DNA testing of women having Papanicolaou (Pap) smears showing atypical squamous cells of undetermined significance (ASCUS) has clinical usefulness. Whether HPV DNA testing alone is useful in primary screening remains to be determined. OBJECTIVE To determine the accuracy of HPV DNA testing for detecting cervical intraepithelial neoplasia (CIN) grade 3 or cancer (the criterion standard). DESIGN, SETTING, AND PARTICIPANTS Between December 1997 and October 2000, 4075 women who attended Planned Parenthood clinics in Washington State were screened simultaneously using thin-layer Pap and HPV DNA testing by a polymerase chain reaction (PCR)-based method and by a liquid-based RNA-DNA hybridization capture with signal amplification assay (signal amplification). Women who were positive for high-risk HPV types, or had Pap results of ASCUS or higher, were considered to have positive screening test results and were referred for colposcopy and biopsy. Additionally, a random sample of women with negative screening test results was referred for colposcopy. Based on individual and combined thin-layer Pap, HPV PCR, and HPV signal amplification test results from the screening and the colposcopy visits, 7 colposcopy triage strategies were defined and evaluated. MAIN OUTCOME MEASURE Sensitivity and specificity for detecting cervical lesions graded CIN 3 or higher for each of the 7 triage strategies. RESULTS The estimated prevalence of CIN 3 or higher was 3.2%. The sensitivity (95% confidence interval) of thin-layer Pap (with a result of > or = ASCUS) for identifying women with CIN 3 or higher was', '2001. Comparison of three management strategies for patients with atypical squamous cells of undetermined significance. Baseline results from a randomized trial. BACKGROUND More than 2 million U.S. women receive an equivocal cervical cytologic diagnosis (atypical squamous cells of undetermined significance [ASCUS]) each year. Effective colposcopy triage strategies are needed to identify the minority of women who have clinically significant disease while avoiding excessive follow-up evaluation for others. METHODS The ASCUS/LSIL (i.e., low-grade squamous intraepithelial lesion) Triage Study (ALTS) is a multicenter, randomized trial comparing the sensitivity and specificity of the following three management strategies to detect cervical intraepithelial neoplasia grade 3 (CIN3): 1) immediate colposcopy (considered to be the reference standard), 2) triage to colposcopy based on human papillomavirus (HPV) results from Hybrid Capture 2(TM) (HC 2) and thin-layer cytology results, or 3) triage based on cytology results alone. This article summarizes the cross-sectional enrollment results for 3488 women with a referral diagnosis of ASCUS. All statistical tests are two-sided. RESULTS Among participants with ASCUS, the underlying prevalence of histologically confirmed CIN3 was 5.1%. Sensitivity to detect CIN3 or above by testing for cancer-associated HPV DNA was 96.3% (95% confidence interval [CI] = 91.6% to 98.8%), with 56.1% of women referred to colposcopy. Sensitivity of a single repeat cytology specimen with a triage threshold of HSIL or above was 44.1% (95% CI = 35.6% to 52.9%), with 6.9% referred. Sensitivity of a lower cytology triage threshold of ASCUS or above was 85.3% (95% CI = 78.2% to 90.8%), with 58.6% referred. CONCLUSIONS HC 2 testing for cancer-associated HPV DNA', 'Baseline cytology, human papillomavirus testing, and risk for cervical neoplasia: A 10-year cohort analysis. As more is learned about the development of cervical cancer, the value of annual Pap smear screening for all women is being questioned. This study was conducted to investigate whether women at higher risk for the development of cervical cancer could be identified by testing for the presence of human papillomavirus (HPV) in the cervical smear. These women could be followed annually, and the interval between screening Pap smears for women at lower risk could be increased. Study participants were women enrolled in the Kaiser Permanente healthcare plan in Portland, Oregon, who underwent annual Pap smear screening between April 1989, and November 1990. More than 20,000 women (20,810 of 23,702) had satisfactory cervical smears with sufficient samples for HPV testing, which was conducted using a polymerase chain reaction-based method with MYO9/11 primers. Most women (83.6%) had at least one follow-up smear during the study period; however, women with atypical squamous cells (ASC) or worse had more smears than women with normal results (mean, 4.4 vs. 3.3). Follow-up was conducted more or less annually for a total period of 122 months. HPV results were not used in deciding patient management. Ninety-six percent of the 20,810 baseline Pap smears were diagnosed as negative (N = 20,156). Thirteen percent of these patients tested positive for HPV. The baseline smears of 654 of the 20,810 women (3.1%) were classified as ASC or worse. Of these 654 smears, 417 (63.8%) were positive for HPV. One hundred seventy-eight women had a cytologic diagnosis of a low-grade squamous intraepithelial lesion or worse; of these, 143 (80.3%) tested positive for HPV. During the 10 years of']	973	27446873	['6561200', '19140422', '16980892', '76415938']	scientific_claim_retrieval	scientific
  0b36f153edb82e0a811376a95d610cc1	mteb/trec-covid	how long can the coronavirus live outside the body	Recommendations for the prevention of transmission of SARS during GI endoscopy	['The impact of respiratory viruses on lung health after preterm birth. Children born preterm, less than 37 weeks’ gestation, are at increased risk of viral respiratory infections and associated complications both during their initial birth hospitalisation and in their first years following discharge. This increased burden of viral respiratory infections is likely to have long term implications for lung health and function in individuals born preterm, particularly those with bronchopulmonary dysplasia. Several hypotheses have been put forward to explain the association between early life viral respiratory infection and development of suboptimal lung health and function later in life following preterm birth. Although preterm infants with diminished lung function, particularly small airways, might be particularly susceptible to asthma and wheezing disorders following viral infection, there is evidence that respiratory viruses can activate number of inflammatory and airway re-modelling pathways. Therefore, the aim of this review is to highlight the perinatal and early life risk factors that may contribute to increased susceptibility to viral respiratory infections among preterm infants during early life and to understand how respiratory viral infection may influence the development of abnormal lung health and function later in life', 'Generation of protective pneumococcal-specific nasal resident memory CD4(+) T cells via parenteral immunization. The generation of tissue-resident memory T cells (T(RM)) is an essential aspect of immunity at mucosal surfaces, and it has been suggested that preferential generation of T(RM) is one of the principal advantages of mucosally administered vaccines. We have previously shown that antigen-specific, IL-17-producing CD4(+) T cells can provide capsular antibody-independent protection against nasal carriage of Streptococcus pneumoniae; but whether pneumococcus-responsive T(RM) are localized within the nasal mucosa and are sufficient for protection from carriage has not been determined. Here we show that intranasal administration of live or killed pneumococci to mice generates pneumococcus-responsive IL-17A-producing CD4(+) mucosal T(RM). Furthermore, we show that these cells are sufficient to mediate long-lived, neutrophil-dependent protection against subsequent pneumococcal nasal challenge. Unexpectedly, and in contrast with the prevailing paradigm, we found that parenteral administration of killed pneumococci also generates protective IL-17A(+)CD4(+) T(RM) in the nasal mucosa. These results demonstrate a critical and sufficient role of T(RM) in prevention of pneumococcal colonization, and further that these cells can be generated by parenteral immunization. Our findings therefore have important implications regarding the generation of immune protection at mucosal surfaces by vaccination', 'Nanorobot Hardware Architecture for Medical Defense. This work presents a new approach with details on the integrated platform and hardware architecture for nanorobots application in epidemic control, which should enable real time in vivo prognosis of biohazard infection. The recent developments in the field of nanoelectronics, with transducers progressively shrinking down to smaller sizes through nanotechnology and carbon nanotubes, are expected to result in innovative biomedical instrumentation possibilities, with new therapies and efficient diagnosis methodologies. The use of integrated systems, smart biosensors, and programmable nanodevices are advancing nanoelectronics, enabling the progressive research and development of molecular machines. It should provide high precision pervasive biomedical monitoring with real time data transmission. The use of nanobioelectronics as embedded systems is the natural pathway towards manufacturing methodology to achieve nanorobot applications out of laboratories sooner as possible. To demonstrate the practical application of medical nanorobotics, a 3D simulation based on clinical data addresses how to integrate communication with nanorobots using RFID, mobile phones, and satellites, applied to long distance ubiquitous surveillance and health monitoring for troops in conflict zones. Therefore, the current model can also be used to prevent and save a population against the case of some targeted epidemic disease', 'Modelling of potentially promising SARS protease inhibitors. In many cases, at the beginning of a high throughput screening experiment some information about active molecules is already available. Active compounds (such as substrate analogues, natural products and inhibitors of related proteins) are often identified in low throughput validation studies on a biochemical target. Sometimes the additional structural information is also available from crystallographic studies on protein and ligand complexes. In addition, the structural or sequence similarity of various protein targets yields a novel possibility for drug discovery. Co-crystallized compounds from homologous proteins can be used to design leads for a new target without co-crystallized ligands. In this paper we evaluate how far such an approach can be used in a real drug campaign, with severe acute respiratory syndrome (SARS) coronavirus providing an example. Our method is able to construct small molecules as plausible inhibitors solely on the basis of the set of ligands from crystallized complexes of a protein target, and other proteins from its structurally homologous family. The accuracy and sensitivity of the method are estimated here by the subsequent use of an electronic high throughput screening flexible docking algorithm. The best performing ligands are then used for a very restrictive similarity search for potential inhibitors of the SARS protease within the million compounds from the Ligand.Info small molecule meta-database. The selected molecules can be passed on for further experimental validation']	15	alo15byi	['am4mdxok', 'dxdr6oan', '49it5pha', 'mw7athol']	biomedical_literature_retrieval	scientific
  9f0131b2d375164750c53d86e4956b28	mteb/scifact	Anthrax spores are very difficult to dispose once they are dispersed.	Secondary aerosolization of viable Bacillus anthracis spores in a contaminated US Senate Office.. CONTEXT Bioterrorist attacks involving letters and mail-handling systems in Washington, DC, resulted in Bacillus anthracis (anthrax) spore contamination in the Hart Senate Office Building and other facilities in the US Capitol's vicinity. OBJECTIVE To provide information about the nature and extent of indoor secondary aerosolization of B anthracis spores. DESIGN Stationary and personal air samples, surface dust, and swab samples were collected under semiquiescent (minimal activities) and then simulated active office conditions to estimate secondary aerosolization of B anthracis spores. Nominal size characteristics, airborne concentrations, and surface contamination of B anthracis particles (colony-forming units) were evaluated. RESULTS Viable B anthracis spores reaerosolized under semiquiescent conditions, with a marked increase in reaerosolization during simulated active office conditions...	['Interchangeability of Biosimilars: A European Perspective. Many of the best-selling ‘blockbuster’ biological medicinal products are, or will soon be, facing competition from similar biological medicinal products (biosimilars) in the EU. Biosimilarity is based on the comparability concept, which has been used successfully for several decades to ensure close similarity of a biological product before and after a manufacturing change. Over the last 10 years, experience with biosimilars has shown that even complex biotechnology-derived proteins can be copied successfully. Most best-selling biologicals are used for chronic treatment. This has triggered intensive discussion on the interchangeability of a biosimilar with its reference product, with the main concern being immunogenicity. We explore the theoretical basis of the presumed risks of switching between a biosimilar and its reference product and the available data on switches. Our conclusion is that a switch between comparable versions of the same active substance approved in accordance with EU legislation is not expected to trigger or enhance immunogenicity. On the basis of current knowledge, it is unlikely and very difficult to substantiate that two products, comparable on a population level, would have different safety or efficacy in individual patients upon a switch. Our conclusion is that biosimilars licensed in the EU are interchangeable', 'Production of cloned mice and ES cells from adult somatic cells by nuclear transfer: how to improve cloning efficiency?. Although it has now been 10 years since the first cloned mammals were generated from somatic cells using nuclear transfer (NT), most cloned embryos usually undergo developmental arrest prior to or soon after implantation, and the success rate for producing live offspring by cloning remains below 5%. The low success rate is believed to be associated with epigenetic errors, including abnormal DNA hypermethylation, but the mechanism of "reprogramming" is unclear. We have been able to develop a stable NT method in the mouse in which donor nuclei are directly injected into the oocyte using a piezo-actuated micromanipulator. Especially in the mouse, only a few laboratories can make clones from adult somatic cells, and cloned mice are never successfully produced from most mouse strains. However, this technique promises to be an important tool for future research in basic biology. For example, NT can be used to generate embryonic stem (NT-ES) cell lines from a patient's own somatic cells. We have shown that NT-ES cells are equivalent to ES cells derived from fertilized embryos and that they can be generated relatively easily from a variety of mouse genotypes and cell types of both sexes, even though it may be more difficult to generate clones directly. In general, NT-ES cell techniques are expected to be applied to regenerative medicine; however, this technique can also be applied to the preservation of genetic resources of mouse strain instead of embryos, oocytes and spermatozoa. This review describes how to improve cloning efficiency and NT-ES cell establishment and further applications', 'Thalassemia and malaria: new insights into an old problem.. The hemoglobinopathies are probably the world's most common genetic diseases: The World Health Organization has estimated that at least 5% of the population are carriers for one or other of the most serious forms, the alpha- and beta-thalassemias and the structural variant hemoglobins S, C, and E, which are found at polymorphic frequencies in many countries. All these hemoglobinopathies are believed to provide protection against malaria, and it is thought that, in malarial regions of the world, natural selection has been responsible for elevating and maintaining their gene frequencies, an idea first proposed 50 years ago by J.B.S. Haldane. Epidemiological studies undertaken in the 1950s on hemoglobin S in Africa provided support for the "malaria hypothesis," but until recently it has proved extremely difficult to verify it for the thalassemias. The application of molecular methods has, however, provided new opportunities to address this old question. Population and molecular genetic analysis of thalassemia variants, and microepidemiological studies of the relationship between alpha-thalassemia and malaria in the southwest Pacific, have provided unequivocal evidence for protection. Surprisingly, some of this protection appears to derive from enhanced susceptibility in very young thalassemic children to both Plasmodium falciparum and, especially, P. vivax, and this early exposure appears to provide the basis for better protection in later life', 'Human cancer cells utilize mitotic DNA synthesis to resist replication stress at telomeres regardless of their telomere maintenance mechanism. Telomeres resemble common fragile sites (CFSs) in that they are difficult-to-replicate and exhibit fragility in mitosis in response to DNA replication stress. At CFSs, this fragility is associated with a delay in the completion of DNA replication until early mitosis, whereupon cells are proposed to switch to a RAD52-dependent form of break-induced replication. Here, we show that this mitotic DNA synthesis (MiDAS) is also a feature of human telomeres. Telomeric MiDAS is not restricted to those telomeres displaying overt fragility, and is a feature of a wide range of cell lines irrespective of whether their telomeres are maintained by telomerase or by the alternative lengthening of telomeres (ALT) mechanism. MiDAS at telomeres requires RAD52, and is mechanistically similar to CFS-associated MiDAS, with the notable exception that telomeric MiDAS does not require the MUS81-EME1 endonuclease. We propose a model whereby replication stress initiates a RAD52-dependent form of break-induced replication that bypasses a requirement for MUS81-EME1 to complete DNA synthesis in mitosis']	114	33872649	['5956016', '22490293', '45218443', '4899981']	scientific_claim_retrieval	scientific

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

Evaluation Dataset

research-embedding-dataset

• Dataset: research-embedding-dataset at 4dc4ce9
• Size: 24 evaluation samples
• Columns: id, source_dataset, query, positive, negatives, query_id, positive_id, negative_ids, task_type, and domain
• Approximate statistics based on the first 24 samples:

  	id	source_dataset	query	positive	negatives	query_id	positive_id	negative_ids	task_type	domain
  type	string	string	string	string	list	string	string	list	string	string
  details	min: 21 tokens mean: 25.5 tokens max: 28 tokens	min: 8 tokens mean: 8.42 tokens max: 10 tokens	min: 12 tokens mean: 19.46 tokens max: 34 tokens	min: 142 tokens mean: 292.88 tokens max: 384 tokens	size: 4 elements	min: 3 tokens mean: 3.46 tokens max: 4 tokens	min: 5 tokens mean: 6.54 tokens max: 9 tokens	size: 4 elements	min: 7 tokens mean: 7.0 tokens max: 7 tokens	min: 3 tokens mean: 3.0 tokens max: 3 tokens

• Samples:

  id	source_dataset	query	positive	negatives	query_id	positive_id	negative_ids	task_type	domain
  7f59e84833d00d3cbb5cc5d83b8b9847	mteb/scifact	Post-transcriptional handling of mitochondrial transcripts occurs in mitochondrial RNA granules.	GRSF1 Regulates RNA Processing in Mitochondrial RNA Granules. Various specialized domains have been described in the cytosol and the nucleus; however, little is known about compartmentalization within the mitochondrial matrix. GRSF1 (G-rich sequence factor 1) is an RNA binding protein that was previously reported to localize in the cytosol. We found that an isoform of GRSF1 accumulates in discrete foci in the mitochondrial matrix. These foci are composed of nascent mitochondrial RNA and also contain RNase P, an enzyme that participates in mitochondrial RNA processing. GRSF1 was found to interact with RNase P and to be required for processing of both classical and tRNA-less RNA precursors. In its absence, cleavage of primary RNA transcripts is abnormal, leading to decreased expression of mitochondrially encoded proteins and mitochondrial dysfunction. Our findings suggest that the foci containing GRSF1 and RNase P correspond to sites where primary RNA transcripts converge to be processed...	['Pseudouridine profiling reveals regulated mRNA pseudouridylation in yeast and human cells. Post-transcriptional modification of RNA nucleosides occurs in all living organisms. Pseudouridine, the most abundant modified nucleoside in non-coding RNAs, enhances the function of transfer RNA and ribosomal RNA by stabilizing the RNA structure. Messenger RNAs were not known to contain pseudouridine, but artificial pseudouridylation dramatically affects mRNA function--it changes the genetic code by facilitating non-canonical base pairing in the ribosome decoding centre. However, without evidence of naturally occurring mRNA pseudouridylation, its physiological relevance was unclear. Here we present a comprehensive analysis of pseudouridylation in Saccharomyces cerevisiae and human RNAs using Pseudo-seq, a genome-wide, single-nucleotide-resolution method for pseudouridine identification. Pseudo-seq accurately identifies known modification sites as well as many novel sites in non-coding RNAs, and reveals hundreds of pseudouridylated sites in mRNAs. Genetic analysis allowed us to assign most of the new modification sites to one of seven conserved pseudouridine synthases, Pus1-4, 6, 7 and 9. Notably, the majority of pseudouridines in mRNA are regulated in response to environmental signals, such as nutrient deprivation in yeast and serum starvation in human cells. These results suggest a mechanism for the rapid and regulated rewiring of the genetic code through inducible mRNA modifications. Our findings reveal unanticipated roles for pseudouridylation and provide a resource for identifying the targets of pseudouridine synthases implicated in human disease', 'Mitochondrial RNA Granules Are Centers for Posttranscriptional RNA Processing and Ribosome Biogenesis.. Cytoplasmic RNA granules play a central role in mRNA metabolism, but the importance of mitochondrial RNA granules remains relatively unexplored. We characterized their proteome and found that they contain a large toolbox of proteins dedicated to RNA metabolism. Investigation of four uncharacterized putative RNA-binding proteins-two RNA helicases, DHX30 and DDX28, and two proteins of the Fas-activated serine-threonine kinase (FASTKD) family, FASTKD2 and FASTKD5-demonstrated that both helicases and FASTKD2 are required for mitochondrial ribosome biogenesis. RNA-sequencing (RNA-seq) analysis showed that DDX28 and FASTKD2 bound the 16S rRNA. FASTKD5 is required for maturing precursor mRNAs that are not flanked by tRNAs and that therefore cannot be processed by the canonical mRNA maturation pathway. Silencing FASTKD5 rendered mature COX I mRNA almost undetectable, which severely reduced the synthesis of COX I, resulting in a complex IV assembly defect. These data demonstrate that mitochondrial RNA granules are centers for posttranscriptional RNA processing and the biogenesis of mitochondrial ribosomes', 'Non-coding RNAs: Functions and applications in endocrine-related cancer.. A significant fraction of the human genome is transcribed as non-coding RNAs (ncRNAs). This non-coding transcriptome has challenged the notion of the central dogma and its involvement in transcriptional and post-transcriptional regulation of gene expression is well established. Interestingly, several ncRNAs are dysregulated in cancer and current non-coding transcriptome research aims to use our increasing knowledge of these ncRNAs for the development of cancer biomarkers and anti-cancer drugs. In endocrine-related cancers, for which survival rates can be relatively low, there is a need for such advancements. In this review, we aimed to summarize the roles and clinical implications of recently discovered ncRNAs, including long ncRNAs, PIWI-interacting RNAs, tRNA- and Y RNA-derived ncRNAs, and small nucleolar RNAs, in endocrine-related cancers affecting both sexes. We focus on recent studies highlighting discoveries in ncRNA biology and expression in cancer, and conclude with a discussion on the challenges and future directions, including clinical application. ncRNAs show great promise as diagnostic tools and therapeutic targets, but further work is necessary to realize the potential of these unconventional transcripts', "Direct isolation and identification of promoters in the human genome.. Transcriptional regulatory elements play essential roles in gene expression during animal development and cellular response to environmental signals, but our knowledge of these regions in the human genome is limited despite the availability of the complete genome sequence. Promoters mark the start of every transcript and are an important class of regulatory elements. A large, complex protein structure known as the pre-initiation complex (PIC) is assembled on all active promoters, and the presence of these proteins distinguishes promoters from other sequences in the genome. Using components of the PIC as tags, we isolated promoters directly from human cells as protein-DNA complexes and identified the resulting DNA sequences using genomic tiling microarrays. Our experiments in four human cell lines uncovered 252 PIC-binding sites in 44 semirandomly selected human genomic regions comprising 1% (30 megabase pairs) of the human genome. Nearly 72% of the identified fragments overlap or immediately flank 5' ends of known cDNA sequences, while the remainder is found in other genomic regions that likely harbor putative promoters of unannotated transcripts. Indeed, molecular analysis of the RNA isolated from one cell line uncovered transcripts initiated from over half of the putative promoter fragments, and transient transfection assays revealed promoter activity for a significant proportion of fragments when they were fused to a luciferase reporter gene. These results demonstrate the specificity of a genome-wide analysis method for mapping transcriptional regulatory elements and also indicate that a small, yet significant number of human genes remains to be discovered"]	962	13931771	['984825', '20018321', '11484808', '25799020']	scientific_claim_retrieval	scientific
  4b94d2486cf7181f8458d2496310f9b2	mteb/trec-covid	how long does coronavirus remain stable on surfaces?	Surface Disinfection. The patient-care areas in a dental setting become contaminated with bacterial and viral pathogens during patient treatment. Incorporating standard precautions set forth by CDC and OSHA guidelines will reduce the risk of disease transmission. Contaminated environmental surfaces, including clinical contact and housekeeping surfaces, become a reservoir of infectious material with the potential to spread an infection to health-care personnel and patients. Transmission of pathogens can occur by direct or indirect contact of clinical contact surfaces and the hands of health-care personnel. Proper infection control protocol of these surfaces includes cleaning, disinfecting, and the use of barriers to prevent the spread of infectious pathogens. This chapter will provide an overview of the disinfection protocol of environmental surfaces in the dental setting. The topics include the various chemical formulations of hospital disinfectants and their proper use, as well as phy...	['Infections and immunological hazards of allogeneic bone transplantation. Allogeneic transplantation of human cancellous and cortical bone is a controversially discussed concept in trauma and orthopaedic surgery. Biological and immunological arguments support transplantation of autologous material whenever this is technically possible. On the other hand, synthetic alloplastic materials for bone substitution are available free of immunological and hygienic hazards. In this context the value of allogeneic bone grafts is discussed, especially considering the problem of AIDS. If autologous corticospongious bone is to be used its supply is limited. On the other hand, alloplastic synthetic artificial bone does not meet all the requirements demanded for substitution of large osseous defects up to now. The problems of geometric and mechanical stability of these alloplastic materials still remain. Therefore, no alternative to allografting of large, stable, corticospongious fragments exists in some cases. Bone transplantation is performed without vital indication in nearly every case. Thus an optimum of hygienic security has to be claimed for recipients of allogeneic bone. The “Munich model” for bone transplantation is presented and discussed', 'Traffic-driven epidemic outbreak on complex networks: How long does it take?. Recent studies have suggested the necessity to incorporate traffic dynamics into the process of epidemic spreading on complex networks, as the former provides support for the latter in many real-world situations. While there are results on the asymptotic scope of the spreading dynamics, the issue of how fast an epidemic outbreak can occur remains outstanding. We observe numerically that the density of the infected nodes exhibits an exponential increase with time initially, rendering definable a characteristic time for the outbreak. We then derive a formula for scale-free networks, which relates this time to parameters characterizing the traffic dynamics and the network structure such as packet-generation rate and betweenness distribution. The validity of the formula is tested numerically. Our study indicates that increasing the average degree and/or inducing traffic congestion can slow down the spreading process significantly', 'A Spatial-Temporal Approach to Differentiate Epidemic Risk Patterns. The purpose of disease mapping is to find spatial clustering and identify risk areas and potential epidemic initiators. Rather than relying on plotting either the case number or incidence rate, this chapter proposes three temporal risk indices: the probability of case occurrence (how often did uneven cases occur), the duration of an epidemic (how long did cases persist), and the intensity of a transmission (were the case of chronological significance). By integrating the three indicators using the local indicator of spatial autocorrelation (LISA) statistic, this chapter intends to develop a novel approach for evaluating spatial-temporal relationships with different risk patterns in the 2002 dengue epidemic, the worst outbreak in the past sixty years. With this approach, not only are hypotheses generated through the mapping processes in furthering investigation, but also procedures provided to identify spatial health risk levels with temporal characteristics', 'Evaluation of TaqMan qPCR System Integrating Two Identically Labelled Hydrolysis Probes in Single Assay. Ongoing evolution of viral pathogens is a significant issue in diagnostic virology employing TaqMan qPCR/RT-qPCR. Specific concerns are related to false negativity due to probe binding failure. One option for compensating for such deficiency is to integrate a second identically labelled probe in the assay. However, how this alteration influences the reaction parameters has not been comprehensively demonstrated. In the present study, we evaluate a TaqMan protocol using two identically labelled hydrolysis probes (simple, LNA (locked-nucleic-acid)) and MGB (minor-groove-binder) modified probes and combinations thereof in a single assay. Our results based on a synthetic amplicon suggest that the second probe does not compromise the TaqMan qPCR/RT-qPCR parameters, which repeatedly and reproducibly remained comparable to those of the corresponding single-probe assays, irrespective of the relative probe orientation, whether opposite or tandem, and probe modifications or combinations thereof. On the other hand, the second probe additively contributed to the overall fluorescence signal. The utility of the dual-probe approach was demonstrated on practical examples by using field specimens. We hope that the present study might serve as a theoretical basis for the development or improvement of TaqMan qPCR/RT-qPCR assays for the detection of highly variable nucleic acid templates']	16	gegboxgs	['tle1j3xv', 'wuy3cwhu', 'q5cmwzr8', 'mne96c9i']	biomedical_literature_retrieval	scientific
  ea7b1574cf9d4329008b832abf0344cb	mteb/scifact	ART substantially reduces infectiveness of HIV-positive people.	Elimination of HIV in South Africa through Expanded Access to Antiretroviral Therapy: A Model Comparison Study. BACKGROUND Expanded access to antiretroviral therapy (ART) using universal test and treat (UTT) has been suggested as a strategy to eliminate HIV in South Africa within 7 y based on an influential mathematical modeling study. However, the underlying deterministic model was criticized widely, and other modeling studies did not always confirm the study's finding. The objective of our study is to better understand the implications of different model structures and assumptions, so as to arrive at the best possible predictions of the long-term impact of UTT and the possibility of elimination of HIV. METHODS AND FINDINGS We developed nine structurally different mathematical models of the South African HIV epidemic in a stepwise approach of increasing complexity and realism. The simplest model resembles the initial deterministic model, while the most comprehensive model is the stoch...	['Universal voluntary HIV testing with immediate antiretroviral therapy as a strategy for elimination of HIV transmission: a mathematical model.. BACKGROUND Roughly 3 million people worldwide were receiving antiretroviral therapy (ART) at the end of 2007, but an estimated 6.7 million were still in need of treatment and a further 2.7 million became infected with HIV in 2007. Prevention efforts might reduce HIV incidence but are unlikely to eliminate this disease. We investigated a theoretical strategy of universal voluntary HIV testing and immediate treatment with ART, and examined the conditions under which the HIV epidemic could be driven towards elimination. METHODS We used mathematical models to explore the effect on the case reproduction number (stochastic model) and long-term dynamics of the HIV epidemic (deterministic transmission model) of testing all people in our test-case community (aged 15 years and older) for HIV every year and starting people on ART immediately after they are diagnosed HIV positive. We used data from South Africa as the test case for a generalised epidemic, and assumed that all HIV transmission was heterosexual. FINDINGS The studied strategy could greatly accelerate the transition from the present endemic phase, in which most adults living with HIV are not receiving ART, to an elimination phase, in which most are on ART, within 5 years. It could reduce HIV incidence and mortality to less than one case per 1000 people per year by 2016, or within 10 years of full implementation of the strategy, and reduce the prevalence of HIV to less than 1% within 50 years. We estimate that in 2032, the yearly cost of the present strategy and the theoretical strategy would both be US$1.7 billion; however, after this time, the cost of the present strategy', "Late presentation of HIV-infected individuals.. Late presentation remains a major concern despite the dramatically improved prognosis realized by ART. We define a first presentation for HIV care during the course of HIV infection as 'late' if an AIDS-defining opportunistic disease is apparent, or if CD4+ T-cells are <200/microl. In the Western world, approximately 10 and 30% of HIV-infected individuals still present with CD4+ T-cells <50 and <200/microl, respectively; estimates are substantially higher for developing countries. Diagnosis and treatment of opportunistic diseases and intense supportive in-hospital care take precedence over ART. Benefits of starting ART without delay, that is, when opportunistic diseases are still active, include faster resolution of opportunistic diseases and a decreased risk of recurrence. The downside of starting ART without delay could include toxicity, drug interactions and immune reconstitution inflammatory syndrome (IRIS). Among asymptomatic or oligosymptomatic individuals presenting late, where ART and primary prophylaxis are initiated, approximately 10-20% will become symptomatic from drug toxicity or undiagnosed opportunistic complications, including IRIS, which require appropriate therapies. In this review we describe late presentation to HIV care, the scale of the problem, the evaluation of a late-presenting patient and challenges associated with initiation of potent antiretroviral therapy (ART) in the setting of acute opportunistic infections and other comorbidities", 'Integrating nutrition security with treatment of people living with HIV: lessons from Kenya.. BACKGROUND The increased caloric requirements of HIV-positive individuals, undesirable side effects of treatment that may be worsened by malnutrition (but alleviated by nutritional support), and associated declines in adherence and possible increased drug resistance are all justifications for developing better interventions to strengthen the nutrition security of individuals receiving antiretroviral treatment. OBJECTIVE To highlight key benefits and challenges relating to interventions aimed at strengthening the nutrition security of people living with HIV who are receiving antiretroviral treatment. METHODS Qualitative research was undertaken on a short-term nutrition intervention linked to the provision of free antiretroviral treatment for people living with HIV in western Kenya in late 2005 and early 2006. RESULTS Patients enrolled in the food program while on treatment regimens self-reported greater adherence to their medication, fewer side effects, and a greater ability to satisfy increased appetite. Most clients self-reported weight gain, recovery of physical strength, and the resumption of labor activities while enrolled in dual (food supplementation and treatment) programs. Such improvements were seen to catalyze increased support from family and community. CONCLUSIONS These findings provide further empirical support to calls for a more holistic and comprehensive response to the coexistence of AIDS epidemics with chronic nutrition insecurity. Future work is needed to clarify ways of bridging the gap between short-term nutritional support to individuals and longer-term livelihood security programming for communities affected by AIDS. Such interdisciplinary research will', 'Prospective Study of One Million Deaths in India: Rationale, Design, and Validation Results. Over 75% of the annual estimated 9.5 million deaths in India occur in the home, and the large majority of these do not have a certified cause. India and other developing countries urgently need reliable quantification of the causes of death. They also need better epidemiological evidence about the relevance of physical (such as blood pressure and obesity), behavioral (such as smoking, alcohol, HIV-1 risk taking, and immunization history), and biological (such as blood lipids and gene polymorphisms) measurements to the development of disease in individuals or disease rates in populations. This report here is on the rationale, design, and implementation of the world’s largest prospective study of the causes and correlates of mortality. Nearly 14 million people in 2.4 million nationally representative Indian households will be monitored (for vital status and, if dead, the causes of death through a well-validated verbal autopsy (VA) instrument. be addressed. This study will reliably document not only the underlying cause of child and adult deaths but also key risk factors (behavioral, physical, environmental, and eventually, genetic). It offers a globally replicable model for reliably estimating cause-specific mortality using VA and strengthens India’s flagship mortality monitoring system. Despite the misclassification that is still expected, the new cause-of-death data will be substantially better than that available previously.[PLoS Medicine, February 2006.]']	61	13901073	['374902', '46353045', '31562330', '17415081']	scientific_claim_retrieval	scientific

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: epoch
• learning_rate: 2e-05
• num_train_epochs: 10
• warmup_steps: 0.1
• fp16: True
• load_best_model_at_end: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• do_predict: False
• eval_strategy: epoch
• prediction_loss_only: True
• per_device_train_batch_size: 8
• per_device_eval_batch_size: 8
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 2e-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
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: None
• warmup_ratio: None
• warmup_steps: 0.1
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• enable_jit_checkpoint: False
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• use_cpu: False
• seed: 42
• data_seed: None
• bf16: False
• fp16: True
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: -1
• ddp_backend: None
• debug: []
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: True
• ignore_data_skip: False
• fsdp: []
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• 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
• 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
• 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_for_metrics: []
• eval_do_concat_batches: True
• auto_find_batch_size: False
• full_determinism: False
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• 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
• use_cache: False
• prompts: None
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Click to expand

Epoch	Step	Training Loss	Validation Loss
0.0658	10	6.4850	-
0.1316	20	0.6843	-
0.1974	30	0.0068	-
0.2632	40	0.0048	-
0.3289	50	0.0007	-
0.3947	60	0.0001	-
0.4605	70	0.0000	-
0.5263	80	0.0000	-
0.5921	90	0.0000	-
0.6579	100	0.0000	-
0.7237	110	0.0000	-
0.7895	120	0.0000	-
0.8553	130	0.0000	-
0.9211	140	0.0000	-
0.9868	150	0.0000	-
1.0	152	-	0.0000
1.0526	160	0.0000	-
1.1184	170	0.0000	-
1.1842	180	0.0000	-
1.25	190	0.0000	-
1.3158	200	0.0000	-
1.3816	210	0.0000	-
1.4474	220	0.0000	-
1.5132	230	0.0000	-
1.5789	240	0.0000	-
1.6447	250	0.0000	-
1.7105	260	0.0000	-
1.7763	270	0.0000	-
1.8421	280	0.0000	-
1.9079	290	0.0000	-
1.9737	300	0.0000	-
2.0	304	-	0.0000
2.0395	310	0.0000	-
2.1053	320	0.0001	-
2.1711	330	0.0000	-
2.2368	340	0.0000	-
2.3026	350	0.0000	-
2.3684	360	0.0000	-
2.4342	370	0.0000	-
2.5	380	0.0000	-
2.5658	390	0.0000	-
2.6316	400	0.0000	-
2.6974	410	0.0000	-
2.7632	420	0.0000	-
2.8289	430	0.0000	-
2.8947	440	0.0000	-
2.9605	450	0.0000	-
3.0	456	-	0.0000
3.0263	460	0.0000	-
3.0921	470	0.0000	-
3.1579	480	0.0000	-
3.2237	490	0.0000	-
3.2895	500	0.0000	-
3.3553	510	0.0000	-
3.4211	520	0.0000	-
3.4868	530	0.0000	-
3.5526	540	0.0000	-
3.6184	550	0.0000	-
3.6842	560	0.0000	-
3.75	570	0.0000	-
3.8158	580	0.0000	-
3.8816	590	0.0000	-
3.9474	600	0.0000	-
4.0	608	-	0.0000
4.0132	610	0.0000	-
4.0789	620	0.0000	-
4.1447	630	0.0000	-
4.2105	640	0.0000	-
4.2763	650	0.0000	-
4.3421	660	0.0000	-
4.4079	670	0.0000	-
4.4737	680	0.0000	-
4.5395	690	0.0000	-
4.6053	700	0.0000	-
4.6711	710	0.0000	-
4.7368	720	0.0000	-
4.8026	730	0.0000	-
4.8684	740	0.0000	-
4.9342	750	0.0000	-
5.0	760	0.0000	0.0000
5.0658	770	0.0000	-
5.1316	780	0.0000	-
5.1974	790	0.0000	-
5.2632	800	0.0000	-
5.3289	810	0.0000	-
5.3947	820	0.0000	-
5.4605	830	0.0000	-
5.5263	840	0.0000	-
5.5921	850	0.0000	-
5.6579	860	0.0000	-
5.7237	870	0.0000	-
5.7895	880	0.0000	-
5.8553	890	0.0000	-
5.9211	900	0.0000	-
5.9868	910	0.0000	-
6.0	912	-	0.0000
6.0526	920	0.0000	-
6.1184	930	0.0000	-
6.1842	940	0.0000	-
6.25	950	0.0	-
6.3158	960	0.0000	-
6.3816	970	0.0000	-
6.4474	980	0.0000	-
6.5132	990	0.0	-
6.5789	1000	0.0000	-
6.6447	1010	0.0000	-
6.7105	1020	0.0000	-
6.7763	1030	0.0000	-
6.8421	1040	0.0000	-
6.9079	1050	0.0000	-
6.9737	1060	0.0000	-
7.0	1064	-	0.0
7.0395	1070	0.0000	-
7.1053	1080	0.0000	-
7.1711	1090	0.0000	-
7.2368	1100	0.0000	-
7.3026	1110	0.0000	-
7.3684	1120	0.0000	-
7.4342	1130	0.0000	-
7.5	1140	0.0000	-
7.5658	1150	0.0000	-
7.6316	1160	0.0000	-
7.6974	1170	0.0000	-
7.7632	1180	0.0000	-
7.8289	1190	0.0000	-
7.8947	1200	0.0000	-
7.9605	1210	0.0000	-
8.0	1216	-	0.0
8.0263	1220	0.0000	-
8.0921	1230	0.0000	-
8.1579	1240	0.0000	-
8.2237	1250	0.0000	-
8.2895	1260	0.0000	-
8.3553	1270	0.0000	-
8.4211	1280	0.0000	-
8.4868	1290	0.0000	-
8.5526	1300	0.0000	-
8.6184	1310	0.0000	-
8.6842	1320	0.0000	-
8.75	1330	0.0000	-
8.8158	1340	0.0000	-
8.8816	1350	0.0000	-
8.9474	1360	0.0000	-
9.0	1368	-	0.0
9.0132	1370	0.0000	-
9.0789	1380	0.0000	-
9.1447	1390	0.0000	-
9.2105	1400	0.0000	-
9.2763	1410	0.0000	-
9.3421	1420	0.0000	-
9.4079	1430	0.0000	-
9.4737	1440	0.0000	-
9.5395	1450	0.0000	-
9.6053	1460	0.0000	-
9.6711	1470	0.0000	-
9.7368	1480	0.0000	-
9.8026	1490	0.0000	-
9.8684	1500	0.0000	-
9.9342	1510	0.0000	-
10.0	1520	0.0000	0.0

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.12.12
• Sentence Transformers: 5.3.0
• Transformers: 5.0.0
• PyTorch: 2.10.0+cu128
• Accelerate: 1.13.0
• Datasets: 4.0.0
• Tokenizers: 0.22.2

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",
}

MultipleNegativesRankingLoss

@misc{oord2019representationlearningcontrastivepredictive,
      title={Representation Learning with Contrastive Predictive Coding},
      author={Aaron van den Oord and Yazhe Li and Oriol Vinyals},
      year={2019},
      eprint={1807.03748},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/1807.03748},
}