|
--- |
|
tags: |
|
- sentence-transformers |
|
- sentence-similarity |
|
- feature-extraction |
|
- generated_from_trainer |
|
- dataset_size:43494 |
|
- loss:TripletLoss |
|
base_model: allenai/specter2_aug2023refresh_base |
|
widget: |
|
- source_sentence: As a result of technological progress, environmental aspects and |
|
social change, the automotive industry is undergoing a radical transformation. |
|
The focus is no longer on the product "vehicle" but much more on the mobility |
|
service itself and the users individual experience and well-being during travel |
|
time. In that field of innovation, the study deals with a explorative investigation |
|
of using the travel time for a improvement of the mental health of the passenger. |
|
The vision is to integrate breathwork relaxation in combination with a human centric |
|
lighting scenario as an immersive service within luxury ride-hailing vehicles |
|
to enhance the mental health during automated rides and utilizing the time spent |
|
in cars for personal pleasure. To enable a user-centered and experimental approach, |
|
a test vehicle from the non-profit company bq.Labs was equipped with the bq breath |
|
work app and a spezialized LED-based lighting screen that was developed by Fraunhofer. |
|
The effects were tested on randomly selected and voluntary users in a guerrilla |
|
testing at three different locations in San Diego. The tests explored user acceptance |
|
of the innovative technologies by combining surveys, vital data collection, qualitative |
|
interviews and observations. Initial data analysis provides insights into the |
|
feasibility and potential effects on well-being and user perception. The study |
|
illustrates those innovations in the field of mobility, involve systemic dependencies |
|
and considerations beyond technology, encompassing social and psychological dimensions. |
|
It underscores that successful innovations require a holistic, user-centered approach |
|
that considers technological, social, and psychological dimensions. The findings |
|
lay the groundwork for future research and development of innovation strategies |
|
in the evolving field of mobility and personalized strength. |
|
sentences: |
|
- 'We examined casual decision-making among a group of participants, which frequently |
|
occurs in daily life. In such a situation, participants do not have strong preferences |
|
for the decision. In addition, because the process of decision-making among people |
|
is part of the time they spend together, it is important to feel enjoyment in |
|
the process and satisfaction with the final decision. In this paper, we propose |
|
a game mechanism for generating a sense of enjoyment in the decision-making process |
|
through communication and a sense of acceptance of the final decision. We experimentally |
|
compared two ways to make decisions about beverages: ) majority voting and ) the |
|
proposed game. In the latter case, the participants enjoyed playing the game and |
|
were satisfied with the decision-making process.' |
|
- This paper presents several important factors affecting the resale prices of used |
|
rental cars. In fact, this paper empirically shows and proves several conjectures |
|
regarding the determinants for used rental car resale values through the use of |
|
detailed micro data from one of the biggest rental car companies. Specifically, |
|
the age of a used car has two composite effects on its resale value, even though |
|
overall the two effects work negatively with a concavity, as rental cars ages. |
|
On the other hand, two mileage variables interact with each other and produce |
|
overall decreasing effects on the resale prices with the opposite interactions. |
|
In terms of the effects of brand image, Hyundai and Renault-Samsung have positive |
|
effects on resale values generally. Ssangyong has a positive effect on the resale |
|
values in the SUV category, and Kia and GM-Daewoo are generally inferior to the |
|
other brands in terms of resale values in all categories. In terms of seasonal |
|
effects, we can conclude that this paper confirms the general perception regarding |
|
seasonal effects on resale values. In details, from November to February, resale |
|
values are affected negatively, and March is the recovering month of increasing |
|
demand in the used car market. August seems to be the highest season for the used |
|
car market due to several demand increases. As a result, this paper plays an important |
|
role in providing a substantial amount of information on the factors affecting |
|
the resale prices of rental cars. |
|
- In this paper we present an approach used to enhance students' competency in software |
|
verification. Students were asked to apply software verification techniques to |
|
a complex formal specification system. The complexity of the system stems from |
|
its sophisticated requirements. Selecting such system for this study was intentional |
|
for the following two reasons ) the system is difficult to understand and analyze |
|
because of the domain knowledge required to generate formal specifications in |
|
temporal logic and ) the system is large and complex which lends itself to a wide |
|
range of applicable verification techniques, and thus highlights the differences |
|
in the capabilities of each of the software verification approaches. Students |
|
were assessed using multiple criteria including; examination in applying learned |
|
techniques, students' attitude toward the technique, perceived efficiency of the |
|
techniques in discovering software defects, and the ability of the technique to |
|
locate errors in the code beyond simply indicating their presence. The results |
|
of this work show that the students applied the learned techniques successfully |
|
and their attitudes towards software verification improved. |
|
- source_sentence: EnglishThe literature has argued that, contrary to what claimed |
|
by the rational economic theory, trade unions have progressively moved towards |
|
the representation of atypical workers by adopting more inclusive strategies of |
|
collective bargaining. The strength and modalities of such strategies are affected |
|
by national institutions of labour market and company-level union representation |
|
to which trade unions can draw in workplaces. Within this context, still remain |
|
to be discovered how the aforementioned institutions are enacted and in what subjects |
|
of employment relations can be used by unions in order to protect atypical work. |
|
This paper deals with these issues. It analyzes how unions have used distinctive |
|
institutional factors with regards of both external and interna! flexibility and |
|
in reference to regular and temporary workers to be able to improve the working |
|
conditions of atypical work. Trade unions negotiated a promotion system to permanent |
|
positions and allowed temporary workers to develop the same skills acquired by |
|
regular employees, which were also beneficial for permanent workers' employment |
|
conditions. The defense of regular workers' employrnent conditions was crucial |
|
in order to maintain an inclusive strategy of collective bargaining. italianoIntroduzione. |
|
- Contesto istituzionale e mobilizazione delle risorse. - Flessibilita ed interessi |
|
della forza del lavoro atipica e regolare. - Disegno e metodo della ricerca. - |
|
La strategia di contrattazione colletiva inclusiva fra flessibilita esterna ed |
|
interna. - Analisi e discussione. - Conclusioni. |
|
sentences: |
|
- Much has been invested in big data and artificial intelligence-based solutions |
|
for healthcare. However, few applications have been implemented in clinical practice. |
|
Early economic evaluations can help to improve decision-making by developers of |
|
analytics underlying these solutions aiming to increase the likelihood of successful |
|
implementation, but recommendations about their use are lacking. The aim of this |
|
study was to develop and apply a framework that positions best practice methods |
|
for economic evaluations alongside development of analytics, thereby enabling |
|
developers to identify barriers to success and to select analytics worth further |
|
investments. |
|
- An in situ field test on nine commonly-used soil water sensors was carried out |
|
in a sandy loam soil located in the Potato Research Center, Fredericton, NB (Canada) |
|
using the gravimetric method as a reference. The results showed that among the |
|
tested sensors, regardless of installation depths and soil water regimes, CS000, |
|
Trase, and Troxler performed the best with the factory calibrations, with a relative |
|
root mean square error (RRMSE) of , , and %, and a r( ) of , , and , respectively. |
|
TRIME, Moisture Point (MP000), and Gopher performed slightly worse with the factory |
|
calibrations, with a RRMSE of , , and %, and a r( ) of , , and , respectively, |
|
while the Gypsum, WaterMark, and Netafim showed a frequent need for calibration |
|
in the application in this region. |
|
- 'The article proposes a comparison between British devolution and Italian one, |
|
both have occurred at about the same time (from the end of Nineties years until |
|
now), looking what is common in devolution process inside two cultural and institutional |
|
context deeply different. About the constitutional innovation, British and Italian |
|
political systems know different method to pass a reform: in British system, Westminster |
|
parliament is sovereign not only in ordinary law-making but above all in constitutional |
|
matter (this is the meaning of parliament sovereignty in Dicey''s thought); in |
|
Italian system, constitutional power isn''t on the same degree of ordinary law, |
|
because the parliament makes ordinary law and an ad hoc convention makes the constitution |
|
(or at least its fundamental reforms), as it''s in French tradition. In spite |
|
of so, it''s possible to see a common element in British and Italian devolution, |
|
on the side of its limits: that is the difficult to compatible the post-centralistic |
|
state and its fiscal autonomy with the universalistic principles of welfare state. |
|
This may be one of the mains challenge that Western states will have to face, |
|
looking for a new political balances for the new era that follows the cold war |
|
end.' |
|
- source_sentence: 'Objective Compared with povidone iodine solution to clean,glutaraldehyde |
|
immersion,highpressure steam sterilization of three disinfection methods for dental |
|
handpiece sterilization effect.Methods Select the dental clinical used over phone,were |
|
randomly divided into A group,B group,C group,D group ,A group for the control |
|
group,only cleaning method did not use any disinfection,B group was % Polyvinylpyrrolidone |
|
iodine solution,wipe,C group were soaked in % glutaraldehyde soluton,D group were |
|
treated with high-pressure steam sterilization,after each phone were inoculated |
|
with bacteria sample dish,the more monitoring of four groups of bacterial culture.Results |
|
A group of bacterial culture for the intensive growth of bacteria,B group had |
|
bacterial growth for the(+ ~+ + +),bacterial growth;C group had bacterial culturegrowth |
|
was(+ ~ + +),bacterial growth;D group of bacterial culture Growth of(-),no bacterial |
|
growth.Conclusion High-pressure steam sterilization was the disinfection of dental |
|
handpieces most effective way. Key words: Disinfection;Dental high-speed equipment' |
|
sentences: |
|
- 'Objective To study the self-locking brackets SmartclipTM 0MXTM MBTTM brackets |
|
and traditional pain comparison.Methods patients with non-extraction orthodontic |
|
treatment were randomly divided into two groups,a group treated with self-locking |
|
brackets,the other group treated with traditional care slot.Patients in orthodontic |
|
treatment of pain within a week were inoestigated by way of a questionnaire survey,including |
|
orthodontic pain,soft tissue irritation,and the strength of a normal life for |
|
patients with the impact.Results Questionnaire response rate was %.The level of |
|
pain was similar in self-ligating bracket group and the traditional bracket group.However,time-related,including |
|
pain after orthodontic treatment was 0h,0 d time,the most intense pain and continued |
|
to 0d,back pain relief,0w about pain relief.Conclusion Self-locking brackets and |
|
brackets have noobvious pain intensity differences,but related with orthodontic |
|
force to the clinical use of force should pay attention to light. Key words: Self-ligating |
|
bracket;Traditional brackets;Orthodontic treatment;Pain' |
|
- Simulation of inflorescences is an important part of virtual plant growth. The |
|
past works about simulation of inflorescences focus mainly on how to generate |
|
inflorescences by However, the productions of L system are difficult to understand |
|
and implement since it is described with rule based language, and especially it |
|
needs too many parameters in simulating inflorescence development and flowering |
|
sequences. Dual scale automaton is a plant growth model based on plant growth |
|
mechanisms, which is easy to understand and implement in programming. In this |
|
paper, the method of simulating inflorescence using dual scale automaton model |
|
is discussed. The dual scale automaton model is improved by introducing the rule |
|
of synchronization development, mechanism of reiteration and delay law of plant |
|
growth from the viewpoints of botany,which make it possible to generate almost |
|
all types of the inflorescences defined by botanists, and to simulate acropetal |
|
and basipetal flowering sequences. Several examples of simulation of typical inflorescences |
|
are given for explaining the theory. The improved model is demonstrated a simpler |
|
but more effective method in simulating inflorescences in comparison with L system. |
|
- 'I N late and through the summer of , the York County Court launched a concerted |
|
attack against Quakers in its part of Massachusetts.* York county magistrate Richard |
|
Waldron arrested three visiting Quaker women and had them beaten out of the jurisdiction; |
|
then, apparently in response to his recommendations, the court proceeded to cite |
|
local Quakers living in Kittery for their failure to attend orthodox church services.l |
|
Although Waldron''s actions did not occur until five years after the General Court |
|
had begun its program of suppressing Quakerism, they seemed consistent with the |
|
general pattern of persecution in seventeenth-century Massachusetts Bay: the discovery |
|
of heterodoxy followed by an immediate attempt to produce local conformity.0 Yet |
|
the considerable delay in Kittery in attacking Quakerism and the lack of any subsequent |
|
systematic effort to produce conformity raise a number of questions regarding |
|
the extent to which prejudice against heterodoxy was the sole motive for suppressing |
|
Quakerism in Massachusetts. Between and the York County Court attacked Quaker |
|
heterodoxy only on a limited number of occasions. Each of the incidents suggests |
|
that Kittery Quakers were punished not because their religious beliefs offended |
|
the court but because those beliefs denoted certain positions on secular issues, |
|
and men like Waldron could employ ecclesiastical sanctions to enlist sources' |
|
- source_sentence: Androgen therapy is the mainstay of treatment for the hypogonadotropic |
|
hypogonadal micropenis because it obviously enhances penis growth in prepubescent |
|
microphallic patients. However, the molecular mechanisms of androgen treatment |
|
leading to penis growth are still largely unknown. To clarify this well-known |
|
phenomenon, we successfully generated a castrated male Sprague Dawley rat model |
|
at puberty followed by testosterone administration. Interestingly, compared with |
|
the control group, testosterone treatment stimulated a dose-dependent increase |
|
of penis weight, length, and width in castrated rats accompanied with a dramatic |
|
recovery of the pathological changes of the penis. Mechanistically, testosterone |
|
administration substantially increased the expression of androgen receptor (AR) |
|
protein. Increased AR protein in the penis could subsequently initiate transcription |
|
of its target genes, including keratin 00B (Krt00b). Importantly, we demonstrated |
|
that KRT00B is generally expressed in the rat penis and that most KRT00B expression |
|
is cytoplasmic. Furthermore, AR could directly modulate its expression by binding |
|
to a putative androgen response element sequence of the Krt00b promoter. Overall, |
|
this study reveals a novel mechanism facilitating penis growth after testosterone |
|
treatment in precastrated prepubescent animals, in which androgen enhances the |
|
expression of AR protein as well as its target genes, such as Krt00b. |
|
sentences: |
|
- This study develops statistical learning models to assess the probability of undergraduate |
|
students graduating within a predetermined period, utilizing admission, performance, |
|
and demographic data. The urgency of addressing student attrition is highlighted |
|
by recent data from the National Center for Education Statistics (NCES), indicating |
|
a % completion rate by full-time undergraduates within six years. This research |
|
leverages institutional data from a Saudi University, focusing on freshmen enrolled |
|
in the - and - academic years, to identify students at risk of dropping out, thereby |
|
enabling timely interventions. Ten algorithms, including decision trees, ensemble |
|
models, SVM, and ANN, were built and evaluated on a test set representing % of |
|
the entire dataset using precision, recall, accuracy, and Matthews correlation |
|
coefficient (MCC). The findings show that SVM and Random Forest models were the |
|
most reliable, achieving accuracies of and respectively, and maintaining balance |
|
in precision, recall, and MCC. Conversely, the naive Bayes model recorded the |
|
worst performance. The comparative analysis revealed the superior performance |
|
of ensemble models over decision tree models in predicting student attrition, |
|
emphasizing the importance of model selection in developing effective early intervention |
|
strategies. In addition, our analysis revealed that academic data is a better |
|
predictor of on-time graduation than admission data, emphasizing the need for |
|
institutions to focus on continuous academic assessment data. |
|
- 'Timely and accurate prediction of human movement in urban areas offers instructive |
|
insights into transportation management, public safety, and location-based services, |
|
to name a few. Yet, modeling urban mobility is challenging and complex because |
|
of the spatiotemporal dynamics of movement behavior and the influence of exogenous |
|
factors such as weather, holidays, and local events. In this paper, we use bus |
|
transportation as a proxy to mine spatiotemporal travel patterns. We propose a |
|
deep-learning-based urban mobility prediction model that collectively forecasts |
|
passenger flows between pairs of city regions in an origin-destination (OD) matrix. |
|
We first process OD matrices in a convolutional neural network to capture spatial |
|
correlations. Intermediate results are reconstructed into three multivariate time |
|
series: hourly, daily, and weekly time series. Each time series is aggregated |
|
in a long short-term memory (LSTM) network with a novel attention mechanism to |
|
guide the aggregation. In addition, our model is context-aware by using contextual |
|
embeddings learned from exogenous factors. We dynamically merge results from LSTM |
|
components and context embeddings in a late fusion network to make a final prediction. |
|
The proposed model is implemented and evaluated using a large-scale transportation |
|
data set of more than million bus trips with a suite of Big Data technologies |
|
developed for data processing. Through performance comparison, we show that our |
|
approach achieves sizable accuracy improvements in urban mobility prediction. |
|
Our work has major implications for efficient transportation system design and |
|
performance improvement. The proposed deep neural network structure is generally |
|
applicable for sequential graph data prediction.' |
|
- Various methods are currently under investigation to preserve fertility in males |
|
treated with high-dose chemotherapy and radiation for malignant and nonmalignant |
|
disorders. Human umbilical cord mesenchymal stem cells (HUC-MSCs), which possess |
|
potent immunosuppressive function and secrete various cytokines and growth factors, |
|
have the potential clinical applications. As a potential alternative, we investigate |
|
whether injection of HUC-MSCs into the interstitial compartment of the testes |
|
to promote spermatogenic regeneration efficiently. HUC-MSCs were isolated from |
|
different sources of umbilical cords and injected into the interstitial space |
|
of one testis from busulfan-treated mice (saline and HEK000 cells injections were |
|
performed in a separate set of mice) and the other testis remained uninjected. |
|
Three weeks after MSCs injection, Relative quantitative reverse transcription |
|
polymerase chain reaction was used to identify the expression of of germ cell |
|
associated, which are all related to meiosis, demonstrated higher levels of spermatogenic |
|
gene expression ( fold) in HUC-MSCs injected testes compared to the contralateral |
|
uninjected testes (five mice). Protein levels for germ cell-specific genes, miwi, |
|
vasa and synaptonemal complex protein (Scp0) were also higher in MSC-treated testes |
|
compared to injected controls weeks after treatment. However, no different expression |
|
was detected in saline water and HEK000 cells injection control group. We have |
|
demonstrated HUC-MSCs could affect mouse germ cell-specific genes expression. |
|
The results also provide a possibility that the transplanted HUC-MSCs may promote |
|
the recovery of spermatogenesis. This study provides further evidence for preclinical |
|
therapeutic effects of HUC-MSCs, and explores a new approach to the treatment |
|
of azoospermia. |
|
- source_sentence: Twenty one surviving infants of pregnancies complicated by rupture |
|
of the membranes during the second trimester that lasted at least one week have |
|
been followed up for a median of months. Five infants ( %) had recurrent respiratory |
|
problems (episodes of wheezing and coughing occurring at least once a week) which |
|
related significantly to the use of neonatal ventilation and to very preterm delivery. |
|
Five of the infants who were born preterm and with birth weights of less than |
|
g had recurrent respiratory symptoms ( %). This compares favourably with an incidence |
|
of symptoms of % among surviving low birthweight infants born at this hospital |
|
after pregnancies not complicated by premature rupture of the membranes. Neither |
|
recurrent respiratory symptoms nor admission to hospital for chest related disorders |
|
were associated with the timing of onset or duration of rupture of the membranes. |
|
We conclude that, among survivors of premature rupture of the membranes, chronic |
|
respiratory morbidity would best be prevented by avoiding very preterm delivery, |
|
regardless of the duration of the rupture. |
|
sentences: |
|
- We report a case of prosthetic valve nocardia endocarditis. A year old farmer |
|
underwent aortic valve replacement with a bioprosthetic valve. The immediate post-operative |
|
course was uneventful but weeks later he developed fever. A trans-oesophageal |
|
echocardiogram (TEE) showed a string like structure attached to the prosthetic |
|
valve. Blood cultures grew N. farcinica. He was initially treated with trimethoprim/sulfamethoxazole |
|
(TMP/SMZ), but due to eosinophilia and leucopenia his treatment was changed to |
|
imipenem and amikacin. He developed a rash, presumed to be due to imipenem, which |
|
was then substituted with linezolid. He completed a week course of intravenous |
|
(i.v.) antibiotics. Desensitization with amoxicillin/clavulanic acid was successful |
|
and the patient received oral amoxicillin/clavulanic acid for months. At present, |
|
months from diagnosis, he is afebrile and TEE is normal. To our knowledge, this |
|
case is the fifth reported case of successful treatment of prosthetic valve nocardia |
|
endocarditis treated without surgery. |
|
- 'The effect of different variants of compiling integrated samples for biochemical |
|
oxygen demand (BOD) kinetics was studied in long-term experiments (up to days) |
|
with water samples taken from the central deep-water region of Lake Onego. It |
|
was a series of experiments carried out simultaneously at and in different seasons |
|
of . Five sampling variants were employed with different horizon combinations: |
|
near surface, near bottom, from different depths in the water column, from the |
|
photic and profundal layers. Two experiments were performed with winter water, |
|
three with summer water, four with autumn water, and seven experiments with spring |
|
water. The most representative sample for studying BOD in long-term experiments |
|
is an sample composed of water from different horizons of the photic layer ( m). |
|
For each variant of integrated sample composition, BOD development in the experiments |
|
was modeled by a corresponding kinetic equation whose parameters represented the |
|
oxidation characteristics of components of the organic matter present in the water |
|
and transformed in the long-term BOD experiment. The resultant kinetic parameters |
|
of BOD were analyzed in relation to the factors determining the final oxidation |
|
of the organic matter components. The patterns in which the type of BOD development |
|
is formed depend on the integrated water sample collection/compilation conditions |
|
and are characterized by the average values of the organic matter contained in |
|
the water, estimated either analytically or from empirical equations, as well |
|
as by the temperature of exposure of water samples in the experiment. Synthesis |
|
of the resultant information showed that the values of BOD kinetic parameters |
|
were generally lower in spring water taken from the central part of Lake Onego |
|
as compared with other seasons, since the oxidation potential of organic matter |
|
components in spring water is higher.' |
|
- Doppler ultrasound measurements of pulmonary blood flow in babies with severe |
|
respiratory distress syndrome treated in a randomised controlled trial of surfactant |
|
replacement showed that the immediate improvement of oxygenation was not associated |
|
with a significant increase in pulmonary blood flow. Reduction in ventilator settings |
|
and increases in the extent of chest wall movements measured by a cardiorespiratory |
|
monitor suggested that the improvement after surfactant had been given was a result |
|
of alveolar stabilisation and increased pulmonary compliance. Further simultaneous |
|
studies of pulmonary blood flow and pulmonary compliance are needed to confirm |
|
these findings. |
|
pipeline_tag: sentence-similarity |
|
library_name: sentence-transformers |
|
metrics: |
|
- cosine_accuracy |
|
model-index: |
|
- name: SentenceTransformer based on allenai/specter2_aug2023refresh_base |
|
results: |
|
- task: |
|
type: triplet |
|
name: Triplet |
|
dataset: |
|
name: discipline tuned specter 2 022 |
|
type: discipline-tuned_specter_2_022 |
|
metrics: |
|
- type: cosine_accuracy |
|
value: 0.9713793103448276 |
|
name: Cosine Accuracy |
|
- task: |
|
type: triplet |
|
name: Triplet |
|
dataset: |
|
name: discipline tuned specter 2 024 |
|
type: discipline-tuned_specter_2_024 |
|
metrics: |
|
- type: cosine_accuracy |
|
value: 0.9710344827586207 |
|
name: Cosine Accuracy |
|
--- |
|
|
|
# SentenceTransformer based on allenai/specter2_aug2023refresh_base |
|
|
|
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [allenai/specter2_aug2023refresh_base](https://huggingface.co/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](https://huggingface.co/allenai/specter2_aug2023refresh_base) <!-- at revision 084e9624d354a1cbc464ef6cc1e3646d236b95d9 --> |
|
- **Maximum Sequence Length:** 512 tokens |
|
- **Output Dimensionality:** 768 dimensions |
|
- **Similarity Function:** Cosine Similarity |
|
<!-- - **Training Dataset:** Unknown --> |
|
<!-- - **Language:** Unknown --> |
|
<!-- - **License:** Unknown --> |
|
|
|
### 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': 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: |
|
|
|
```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("m7n/discipline-tuned_specter_2_024") |
|
# Run inference |
|
sentences = [ |
|
'Twenty one surviving infants of pregnancies complicated by rupture of the membranes during the second trimester that lasted at least one week have been followed up for a median of months. Five infants ( %) had recurrent respiratory problems (episodes of wheezing and coughing occurring at least once a week) which related significantly to the use of neonatal ventilation and to very preterm delivery. Five of the infants who were born preterm and with birth weights of less than g had recurrent respiratory symptoms ( %). This compares favourably with an incidence of symptoms of % among surviving low birthweight infants born at this hospital after pregnancies not complicated by premature rupture of the membranes. Neither recurrent respiratory symptoms nor admission to hospital for chest related disorders were associated with the timing of onset or duration of rupture of the membranes. We conclude that, among survivors of premature rupture of the membranes, chronic respiratory morbidity would best be prevented by avoiding very preterm delivery, regardless of the duration of the rupture.', |
|
'Doppler ultrasound measurements of pulmonary blood flow in babies with severe respiratory distress syndrome treated in a randomised controlled trial of surfactant replacement showed that the immediate improvement of oxygenation was not associated with a significant increase in pulmonary blood flow. Reduction in ventilator settings and increases in the extent of chest wall movements measured by a cardiorespiratory monitor suggested that the improvement after surfactant had been given was a result of alveolar stabilisation and increased pulmonary compliance. Further simultaneous studies of pulmonary blood flow and pulmonary compliance are needed to confirm these findings.', |
|
'The effect of different variants of compiling integrated samples for biochemical oxygen demand (BOD) kinetics was studied in long-term experiments (up to days) with water samples taken from the central deep-water region of Lake Onego. It was a series of experiments carried out simultaneously at and in different seasons of . Five sampling variants were employed with different horizon combinations: near surface, near bottom, from different depths in the water column, from the photic and profundal layers. Two experiments were performed with winter water, three with summer water, four with autumn water, and seven experiments with spring water. The most representative sample for studying BOD in long-term experiments is an sample composed of water from different horizons of the photic layer ( m). For each variant of integrated sample composition, BOD development in the experiments was modeled by a corresponding kinetic equation whose parameters represented the oxidation characteristics of components of the organic matter present in the water and transformed in the long-term BOD experiment. The resultant kinetic parameters of BOD were analyzed in relation to the factors determining the final oxidation of the organic matter components. The patterns in which the type of BOD development is formed depend on the integrated water sample collection/compilation conditions and are characterized by the average values of the organic matter contained in the water, estimated either analytically or from empirical equations, as well as by the temperature of exposure of water samples in the experiment. Synthesis of the resultant information showed that the values of BOD kinetic parameters were generally lower in spring water taken from the central part of Lake Onego as compared with other seasons, since the oxidation potential of organic matter components in spring water is higher.', |
|
] |
|
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] |
|
``` |
|
|
|
<!-- |
|
### Direct Usage (Transformers) |
|
|
|
<details><summary>Click to see the direct usage in Transformers</summary> |
|
|
|
</details> |
|
--> |
|
|
|
<!-- |
|
### Downstream Usage (Sentence Transformers) |
|
|
|
You can finetune this model on your own dataset. |
|
|
|
<details><summary>Click to expand</summary> |
|
|
|
</details> |
|
--> |
|
|
|
<!-- |
|
### Out-of-Scope Use |
|
|
|
*List how the model may foreseeably be misused and address what users ought not to do with the model.* |
|
--> |
|
|
|
## Evaluation |
|
|
|
### Metrics |
|
|
|
#### Triplet |
|
|
|
* Datasets: `discipline-tuned_specter_2_022` and `discipline-tuned_specter_2_024` |
|
* Evaluated with [<code>TripletEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.TripletEvaluator) |
|
|
|
| Metric | discipline-tuned_specter_2_022 | discipline-tuned_specter_2_024 | |
|
|:--------------------|:-------------------------------|:-------------------------------| |
|
| **cosine_accuracy** | **0.9714** | **0.971** | |
|
|
|
<!-- |
|
## Bias, Risks and Limitations |
|
|
|
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.* |
|
--> |
|
|
|
<!-- |
|
### Recommendations |
|
|
|
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.* |
|
--> |
|
|
|
## Training Details |
|
|
|
### Training Dataset |
|
|
|
#### Unnamed Dataset |
|
|
|
|
|
* Size: 43,494 training samples |
|
* Columns: <code>anchor</code>, <code>positive</code>, and <code>negative</code> |
|
* Approximate statistics based on the first 1000 samples: |
|
| | anchor | positive | negative | |
|
|:--------|:-------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------| |
|
| type | string | string | string | |
|
| details | <ul><li>min: 80 tokens</li><li>mean: 232.53 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 81 tokens</li><li>mean: 230.16 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 86 tokens</li><li>mean: 229.66 tokens</li><li>max: 512 tokens</li></ul> | |
|
* Samples: |
|
| anchor | positive | negative | |
|
|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| |
|
| <code>Lupus nephritis (LN) is one of the major risk factors for morbidity and overall mortality in systemic lupus erythematosus (SLE). Its pathogenesis is multifactorial, and a number of risk factors, including serological markers, have been identified in recent years, correlating with clinical course and disease severity. Furthermore, a distinctive autoantibody profile has recently been reported in African- American SLE women with LN. The aim of this study was to characterize the autoantibody profile in African-American SLE patients, with LN and without. Only anti-dsDNA achieved statistical significance between the two groups (P < ). Fourteen ( %) patients with LN and ( %) without it exhibited positive anti-Ro/SS-A, anti-Sm, and anti-nRNP, but without anti-La/SS B (P > ). We conclude that African-American SLE patients with LN do not exhibit a specific or distinctive autoantibody profile. However, our data confirm the value of anti-dsDNA in SLE patients with LN.</code> | <code>TRIM00 is a member of the tripartite motif family proteins and is one of the autoantigens which react with anti-SS-A antibody (Ab) present in sera of patients with systemic lupus erythematosus (SLE) and Sjogren's syndrome. Previous studies have shown that TRIM00 dysfunction promotes aberrant B-cell differentiation and Ab production in SLE, and anti-TRIM00 Ab may be related to the TRIM00 dysfunction in human SLE pathogenesis. Here, we examined the relationship between anti-TRIM00 Ab and clinical and immunological characteristics in SLE patients.Twenty-seven patients with SLE ( women and four men) before immunosuppressive therapies, who fulfilled the revised American College of Rheumatology criteria for SLE, and four healthy controls ( women and one man) were enrolled in the study. SLE patients were divided into two groups according to the seropositivity for anti-TRIM00 Ab. Serum anti-TRIM00 Ab levels were measured using enzyme-linked immunosorbent assays. The serum levels of cytokines a...</code> | <code>We construct a stochastic model of real estate pricing. The method of the pricing construction is based on a sequential comparison of the supply prices. We proof that under standard assumptions imposed upon the comparison coefficients there exists an unique non-degenerated limit in distribution and this limit has the lognormal law of distribution. The accordance of empirical distributions of prices to thetheoretically obtained log-normal distribution we verify by numerous statistical data of real estate prices from Saint-Petersburg (Russia). For establishing this accordance we essentially apply the efficient and sensitive test of fit of Kolmogorov-Smirnov. Basing on "The Russian Federal Estimation Standard N0", we conclude that the most probable price, i.e. mode of distribution, is correctly and uniquely defined under the log-normal approximation. Since the mean value of log-normal distribution exceeds the mode - most probable value, it follows that the prices valued by the mathematica...</code> | |
|
| <code>A laboratory prototype of an enzyme biosensor based on pHsensitive field-effect transistors has been developed to determine the total content of indole alkaloids in Rauwolfia serpentina Benth. Ex Kurz tissue culture. The biosensor was characterized by high sensitivity to th A laboratory prototype of an enzyme biosensor based on pHsensitive field effect transistors has been developed to determine the total content of indole alkaloids in Rauwolfia serpentina Benth. Ex Kurz tissue culture. The biosensor was characterized by high sensitivity to the total content of indole alkaloids (minimum limit of determination g/ml of the total content of indole alkaloids contained in the juice obtained from tissue culture of Rauwolfia serpentina). The linear range of biosensor determination of the analyte was from to g / ml of the total content of indole alkaloids. Analysis of indole alkaloids using a biosensor is simple and fast and does not require expensive equipment and special sample preparation f...</code> | <code>A procedure of separate biosensor analysis of the multicomponent sample with aflatoxins and pesticides has been developed and optimized. Biosensor determination of aflatoxins and pesticides was performed using enzyme inhibition analysis. For creation of bioselective element we used enzyme acetylcholinesterase which is co-immobilized with bovine serum albumin on the surface of potentiometric transducer by glutaraldehyde covalent crosslinking. As transducers were pH-sensitive field effect transistors. The concentration of acetylcholine chloride as a substrate for subsequent inhibition analysis was fit; optimal time of inhibition by toxins solution was determinate together with concentration of reactivator (pyridine- -aldoxymmethyliodyd) and time of enzyme reactivation after inhibition. A synergism between trichlorfon and aflatoxin B0 in inhibition of immobilized on a surface pH-sensitive field-effect transistors acetylcholinesterase was investigated. The proposed procedure allows selecti...</code> | <code>Objective: To observe the effect of modified Zhenwu decoction on blood glucose and blood lipid of experimental diabetic rats.Methods: Diabetic model rats randomly were divided into normal control group,diabetic modeling group,modified Zhenwu decoction group.Establish intraperitoneal injection of Streptozotocin diabetic animal models by,after eight weeks blood glucose and blood lipids were detrmined.Results: After the treatment by modified Zhenwu decoction,blood glucose,blood lipid and other indicators improved significantly.Conclusion: Modified Zhenwu decotion can improve the level of renal lower blood glucose and lipid in diabetic rats.</code> | |
|
| <code>In two successive years ( and ), a set of commercial sugar beet cultivars was established in Randomized Complete Block experiments at two sites in central Greece. Cultivar combination was different between years, but not between sites. Leaf sampling took place once during the growing season and leaf area, LA [cm0], leaf midvein length, L [cm] and maximum leaf width, W [cm] were determined using an image analysis system. Leaf parameters were mainly affected by cultivars. Leaf dimensions and their squares (L0, W0) did not provide an accurate model for LA predictions. Using LW as an independent variable, a quadratic model (y = x0 - x + , r = , p< , n = ) provided the most accurate estimation of LA. With compromises in accuracy, the linear relationship between LW and LA (y = x + , r = , p< , n = ) could be used as a prediction model thanks to its simplicity.</code> | <code>The general increase in temperature, together with sudden episodes of extreme temperatures, are increasingly impacting plant species in the present climate change scenario. Limoniastrum monopetalum is a halophyte from the Mediterranean Basin, exposed to broad daily and seasonal changes in temperature and extreme high temperatures. We studied the photosynthetic responses (chlorophyll fluorescence dynamics and gas exchange) of L. monopetalum leaves exposed to temperatures from .0C to .0C under darkness in controlled laboratory conditions. L. monopetalum presented its optimum temperature for photosynthesis around +00C. The photosynthetic apparatus of L. monopetalum exhibited permanent damages at > .0C. L. monopetalum tolerated, without permanent damages, temperatures as low as .0C in darkness. L. monopetalum appears as a plant species very well adapted to the seasonality of the Mediterranean climate, which may work as a pre-adaptation to stand more extreme temperatures in the actual conte...</code> | <code>The article depicts direct and hidden (implicit and explicit) information giving in advertisement discourse, meaning advertising slogans. Having investigated this topic thoroughly, the author found out that cognitive types of presupposition and communicative implicatures played a great role in advertising slogans. There are definitions of phenomena "implicit" and "explicit" with examples. The cognitive types of presupposition (semantic and pragmatic) and their typology is discussed in the article. There is a possibility to figure out what strategy of communicative influence on human's cognition is. Some laws of neurolinguistic programming is also discussed.</code> | |
|
* Loss: [<code>TripletLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#tripletloss) with these parameters: |
|
```json |
|
{ |
|
"distance_metric": "TripletDistanceMetric.COSINE", |
|
"triplet_margin": 0.4 |
|
} |
|
``` |
|
|
|
### Evaluation Dataset |
|
|
|
#### Unnamed Dataset |
|
|
|
|
|
* Size: 2,174 evaluation samples |
|
* Columns: <code>anchor</code>, <code>positive</code>, and <code>negative</code> |
|
* Approximate statistics based on the first 1000 samples: |
|
| | anchor | positive | negative | |
|
|:--------|:-------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------| |
|
| type | string | string | string | |
|
| details | <ul><li>min: 83 tokens</li><li>mean: 235.71 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 82 tokens</li><li>mean: 234.64 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 86 tokens</li><li>mean: 225.92 tokens</li><li>max: 512 tokens</li></ul> | |
|
* Samples: |
|
| anchor | positive | negative | |
|
|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| |
|
| <code>In Organic Law / of 0rd October of the general arrangement of the educational system (LOGSE), the educational system includes the general regime education and the special regime education. Dance is included in the special regime as part of the artistic disciplines together with music, drama, the plastic arts and design. The aim of this article is to analyse the treatment given to Dance in the general regime. Thus, we will try to emphasize the inconsistency that exists between the areas of primary education, which will be obligatory and will have a global and integrated character, and the training of future teachers.</code> | <code>This work aims to analyze the treatment of health education in school textbooks during the period , and to compare it with the one that is conducted at present. It will attempt to verify how many current concepts and ideas were already present in those decades. In addition, the differences in the way of carrying out health education then and now will be outlined, especially those referred to pedagogic strategies and didactic materials. All this will be done from a double perspective: . The concept of health, hygiene and pedagogy of health education. . The program contents of health education in the didactic materials.</code> | <code>The vane-in-cup (VIC) geometry has been widely used for the rheological characterization of yield-stress fluids because it minimizes slip effects at the liquid/solid interface of the rotating geometry and reduces sample damage during the loading process. However, severe kinematic limitations arising from the spatial complexity of mixed shear and extensional flow have been identified for quantitative rheometrical measurements in complex fluids. Recently, vanes with fractal cross sections have been suggested as alternatives for accurate rheometry of elastoviscoplastic fluids. In this work, the steady fractal vane-in-cup (fVIC) flow of a Newtonian fluid and a nonthixotropic Carbopol®️ microgel as well as the unsteady flow of a thixotropic -Carrageenan gel are analyzed using rheo-particle image velocimetry (Rheo-PIV). We describe the velocity distributions in all cases and show that the fVIC produces an almost axisymmetric flow field and rotation rate-independent "effective radius" when us...</code> | |
|
| <code>An ultrahigh vacuum three-axis cryogenic sample manipulator suitable for angle-resolved photoelectron spectroscopy experiments was developed. The sample manipulator is constructed by combining three modules with translation, polar rotation, and azimuthal-tilt rotation capabilities. Polar rotation and the azimuthal-tilt rotation are performed using a differentially pumped rotary stage and a sample goniometer, respectively. Continuous rotation around the polar axis is possible. The sample goniometer is capable of azimuthal rotation of up to and tilt rotation from to , measured from the plane normal to the polar axis. Nonmagnetic materials are used near the sample holder of the goniometer. The sample holder can be cooled using a continuous-flow cryostat. To serve as a radiation shield, the lower portion of the goniometer surrounding the sample holder is cooled separately by another cell filled with liquid nitrogen. With liquid nitrogen or liquid helium for the cryostat, the sample holder ...</code> | <code>In the soft x-ray region below keV, various electron yield (EY) techniques have been employed in x-ray absorption fine structure (XAFS) measurements of bulk materials. The fluorescent x-ray yield (FY) is also utilized for samples of low concentration. Although FY becomes much smaller for lighter elements, it has several advantages compared with EY to measure XAFS spectra; for example, a higher signal-to-background ratio and applicability to insulating materials. However, it has been thought to be unsuitable for concentrated samples due to a self-absorption effect. In this report, the sampling depth and self-absorption effect for bulk concentrated samples are discussed concerning XAFS measurements in a few keV energy region. Some typical FY XAFS spectra of concentrated materials, including insulators, are presented.</code> | <code>To investigate the distribution characteristics of TCM syndromes and the related herbal prescriptions for malignant tumors (MT). A clinical database of the TCM syndromes and the herbal prescriptions in treatment of MT patients were established. The data were then analyzed using cluster and frequency analysis. According to the cluster analysis, the TCM syndromes in MT patients mainly included two patterns: deficiency of both Qi and Yin and internal accumulation of toxic heat. The commonly-prescribed herbs were Huangqi (Astraglus), Nuzhenzi (Fructus Ligustri Lucidi), Lingzhi (Ganoderma Lucidum), Huaishan (Dioscorea Opposita), Xiakucao (Prunella Vulgaris), and Baihuasheshecao (Herba Hedyotidis). Deficiency of Qi and Yin is the primary syndrome of MT, and internal accumulation of toxic heat is the secondary syndrome. The herbs for Qi supplementation and Yin nourishment are mainly used, with the assistance of herbs for heat-clearance and detoxification.</code> | |
|
| <code>Abstract Abstract Worldwide opposition to different aspects of globalisation indicates the emergence of a global social movement that typically targets the international bodies that regulate global trade and global finance, as well as the regulations themselves. The significance of the movement calls for a synthetic analysis that moves beyond the currently used fragmentary descriptions. A more profound conceptual framework will enable researchers to better understand the full dynamic of the movement within its global context In this article we explore the possibilities of applying David Korten's ideal-typical notion of fourth generation development to the anti-globalisation movement. We ask whether anti-globalisation organisation exhibits so-called Fourth Generation characteristics and activities. Our goal is to determine the extent to which the movement as a whole, and the individual organisations which constitute it, conform to the fourth generation development conceptual framework. ...</code> | <code>Abstract Globalisation is a complex, multi-faceted, phenomenon with widely contested meanings. While it has roots in the history of colonialism, capitalist development and imperialism, there are strong indications that what we are witnessing, since the 0000s, is a qualitative break with the past. Old boundaries, categories and meanings are being challenged in profound ways. New forms of exploitation and subjugation emerge in such a way that stark brutal force coexists with and may be increasingly supplanted by more subtle, pervasive forces of hegemonic rule. The latter, however, has opened up new terrains of struggle for people, movements, and governments opposed to one-dimensional 'corporate globalisation', seeking instead the globalisation of social and environmental justice. A continent like Africa much of which has sunk deeper into a 'fourth world' status of extreme under-development, social instability and neo-colonial dependence faces stark choices. Does it seek to partially or f...</code> | <code>So much has been written about the nation vis-a-vis other fields in the humanities, literature in particular. My interest in dance lies in its peculiar location within and vis-a-vis the discourse of the nation. An ephemeral form, dance has elicited various, and even contradictory, valuations; most of the time it is considered a mere form of entertainment. It is undeniable, though, that dance has articulated and informed our ideas of the nation and nationhood. In this paper, I explore how three contemporary dance companies based in Quezon City (The University of the Philippines Dance Company, Airdance, and Dance Forum) have rendered their imaginings of the Philippine nation. I focus on Philippine contemporary dance because as a cultural practice, I believe that it has choreographed the many trajectories and issues embodied in the Philippines's imagining of itself. A number of choreographies by the three companies mobilize motifs, forms, structures, and styles that constitute and signify...</code> | |
|
* Loss: [<code>TripletLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#tripletloss) with these parameters: |
|
```json |
|
{ |
|
"distance_metric": "TripletDistanceMetric.COSINE", |
|
"triplet_margin": 0.4 |
|
} |
|
``` |
|
|
|
### Training Hyperparameters |
|
#### Non-Default Hyperparameters |
|
|
|
- `eval_strategy`: steps |
|
- `per_device_train_batch_size`: 4 |
|
- `per_device_eval_batch_size`: 32 |
|
- `learning_rate`: 7e-06 |
|
- `weight_decay`: 0.01 |
|
- `num_train_epochs`: 1 |
|
- `warmup_ratio`: 0.5 |
|
- `fp16`: True |
|
- `batch_sampler`: no_duplicates |
|
|
|
#### All Hyperparameters |
|
<details><summary>Click to expand</summary> |
|
|
|
- `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`: 32 |
|
- `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`: 7e-06 |
|
- `weight_decay`: 0.01 |
|
- `adam_beta1`: 0.9 |
|
- `adam_beta2`: 0.999 |
|
- `adam_epsilon`: 1e-08 |
|
- `max_grad_norm`: 1.0 |
|
- `num_train_epochs`: 1 |
|
- `max_steps`: -1 |
|
- `lr_scheduler_type`: linear |
|
- `lr_scheduler_kwargs`: {} |
|
- `warmup_ratio`: 0.5 |
|
- `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`: 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 |
|
|
|
</details> |
|
|
|
### Training Logs |
|
| Epoch | Step | Training Loss | Validation Loss | discipline-tuned_specter_2_022_cosine_accuracy | discipline-tuned_specter_2_024_cosine_accuracy | |
|
|:------:|:----:|:-------------:|:---------------:|:----------------------------------------------:|:----------------------------------------------:| |
|
| 0.0023 | 25 | 0.2976 | 0.2980 | 0.9518 | - | |
|
| 0.0046 | 50 | 0.3008 | 0.2969 | 0.9518 | - | |
|
| 0.0069 | 75 | 0.3088 | 0.2953 | 0.9524 | - | |
|
| 0.0092 | 100 | 0.3047 | 0.2929 | 0.9530 | - | |
|
| 0.0115 | 125 | 0.2879 | 0.2897 | 0.9530 | - | |
|
| 0.0138 | 150 | 0.2705 | 0.2855 | 0.9532 | - | |
|
| 0.0161 | 175 | 0.2771 | 0.2804 | 0.9536 | - | |
|
| 0.0184 | 200 | 0.2737 | 0.2744 | 0.9548 | - | |
|
| 0.0207 | 225 | 0.2737 | 0.2676 | 0.9553 | - | |
|
| 0.0230 | 250 | 0.2569 | 0.2600 | 0.9557 | - | |
|
| 0.0253 | 275 | 0.2518 | 0.2512 | 0.9579 | - | |
|
| 0.0276 | 300 | 0.2445 | 0.2416 | 0.9580 | - | |
|
| 0.0299 | 325 | 0.2214 | 0.2310 | 0.9591 | - | |
|
| 0.0322 | 350 | 0.2359 | 0.2204 | 0.9606 | - | |
|
| 0.0345 | 375 | 0.2072 | 0.2090 | 0.9615 | - | |
|
| 0.0368 | 400 | 0.1907 | 0.1976 | 0.9618 | - | |
|
| 0.0391 | 425 | 0.1881 | 0.1850 | 0.9624 | - | |
|
| 0.0414 | 450 | 0.1842 | 0.1733 | 0.9637 | - | |
|
| 0.0437 | 475 | 0.1618 | 0.1628 | 0.9646 | - | |
|
| 0.0460 | 500 | 0.1638 | 0.1533 | 0.9645 | - | |
|
| 0.0483 | 525 | 0.1569 | 0.1440 | 0.9648 | - | |
|
| 0.0506 | 550 | 0.1473 | 0.1354 | 0.9657 | - | |
|
| 0.0529 | 575 | 0.1333 | 0.1281 | 0.9671 | - | |
|
| 0.0552 | 600 | 0.1481 | 0.1223 | 0.9671 | - | |
|
| 0.0575 | 625 | 0.1263 | 0.1167 | 0.9675 | - | |
|
| 0.0598 | 650 | 0.114 | 0.1120 | 0.9684 | - | |
|
| 0.0621 | 675 | 0.1097 | 0.1081 | 0.9693 | - | |
|
| 0.0644 | 700 | 0.1152 | 0.1044 | 0.9698 | - | |
|
| 0.0667 | 725 | 0.1009 | 0.0999 | 0.9705 | - | |
|
| 0.0690 | 750 | 0.0895 | 0.0961 | 0.9709 | - | |
|
| 0.0713 | 775 | 0.0855 | 0.0934 | 0.9711 | - | |
|
| 0.0736 | 800 | 0.0853 | 0.0912 | 0.9715 | - | |
|
| 0.0759 | 825 | 0.0942 | 0.0885 | 0.9714 | - | |
|
| 0.0782 | 850 | 0.1035 | - | - | 0.9710 | |
|
|
|
|
|
### Framework Versions |
|
- Python: 3.10.12 |
|
- Sentence Transformers: 3.3.1 |
|
- Transformers: 4.49.0.dev0 |
|
- PyTorch: 2.5.1+cu121 |
|
- Accelerate: 1.2.1 |
|
- Datasets: 3.2.0 |
|
- Tokenizers: 0.21.0 |
|
|
|
## Citation |
|
|
|
### BibTeX |
|
|
|
#### Sentence Transformers |
|
```bibtex |
|
@inproceedings{reimers-2019-sentence-bert, |
|
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks", |
|
author = "Reimers, Nils and Gurevych, Iryna", |
|
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing", |
|
month = "11", |
|
year = "2019", |
|
publisher = "Association for Computational Linguistics", |
|
url = "https://arxiv.org/abs/1908.10084", |
|
} |
|
``` |
|
|
|
#### TripletLoss |
|
```bibtex |
|
@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} |
|
} |
|
``` |
|
|
|
<!-- |
|
## Glossary |
|
|
|
*Clearly define terms in order to be accessible across audiences.* |
|
--> |
|
|
|
<!-- |
|
## Model Card Authors |
|
|
|
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.* |
|
--> |
|
|
|
<!-- |
|
## Model Card Contact |
|
|
|
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.* |
|
--> |
