README.md

metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:650596
  - loss:CachedGISTEmbedLoss
base_model: BAAI/bge-small-en-v1.5
widget:
  - source_sentence: >-
      Represent this sentence for searching relevant passages: How does a
      high-carbohydrate diet affect inflammation markers and cytokine levels in
      goats?
    sentences:
      - >-
        During carcinogenesis, the tested lactobacilli mix, especially the
        anti-inflammatory M2-programming VD23 strain, ameliorates the
        inflammatory conditions (in the early stages) and/or the
        pro-inflammatory M1-programming MS3 strain can boost an anti-tumour
        immune response with the down-stream effect of eliminating dysplastic
        and cancerous cells. With respect to long-term study of CRC, where
        cancer arises from chronic inflammation and leads to an
        immunosuppressive state with tumour presence, a mixture of probiotic
        bacteria with both anti- and pro-inflammatory (M2- and M1-programming)
        features was used, and this may represent a realistic approach to
        harnessing probiotic strains in the modulation of CRC. 

        While body weight gain over the experimental period did not differ,
        there was a significant difference in daily food intake between all
        experimental groups. Despite the increased food intake of the DMH group
        compared to the DMH+P group, the rats’ ability to convert food into body
        mass (expressed by FER) was not significantly affected. The
        probiotic-fed group was shown to have the highest FER, therefore it can
        be suggested that probiotic treatment can improve absorption and
        digestion of food.
      - "LBP is highly sensitive to LPS, and its plasma levels drastically raise up to 200% in goats fed HC diets and hence considered as reliable biomarker of systemic inflammation (Chang, Zhang, Xu, Jin, Seyfert, et al.,\_; Dong et al.,\_). The APPs production is stimulated by HC diet‐derived LPS in liver through activation of toll‐like receptor‐4 (TLR‐4)‐mediated nuclear factor kappa B (NF‐kB)‐tumor necrosis factor‐α (TNF‐α) signaling pathway in immune cells (Ciesielska et al.,\_; Kany et al.,\_). It has been shown that HC diets induce NF‐κB expression through LPS and thereby modulate the expressions of related cytokines, such as TNF‐α, interleukin‐1β (IL‐1β), IL‐6, and IL‐10, and consequently altered the AAPs production in livers of ruminants (Chang, Zhang, Xu, Jin, Guo, et al.,\_; Dong et al.,\_; Guo et al.,\_)."
      - "After 48\_h transfection, cells were used in the electrophysiology assays in the automated whole-cell patch clamp system QPatch 16X (Sophion Bioscience). \nThe extracellular solution comprised 140 NaCl, 5 KCl, 10 CaCl 2, 2 MgCl 2, 10 glucose and 10 HEPES at pH 7.4 and 320 mOsm. The intracellular solution comprised (in mM) 150 KCl, 1 MgCl 2, 4 NaCl, 0.5 EGTA and 10 HEPES at pH 7.4 and 320 mOsm. Cells were maintained at a holding potential –90\_mV and K + currents elicited by +20\_mV pulse for 500\_ms followed by –40\_mV pulse for additional 500\_ms."
  - source_sentence: >-
      Represent this sentence for searching relevant passages: What software is
      used for carrying out statistics in experiments?
    sentences:
      - >-
        Regarding to the association of dietary intake and CRC, the cases with
        TT genotype of FTO rs9939609 polymorphism had lower intake of copper
        (1.49 ± 0.64 vs. 1.76 ± 0.71 g/d, p = 0.02), selenium (56.15 ± 22.97 vs.
        67.26 ± 15.11 g/d, p < 0.01), β-carotene (2189.73 ± 474.3 vs.
        2461.75 ± 772.57 g/d, p = 0.01), vitamin E (10.58 ± 4.14 vs.
        13.99 ± 6.4 g/d, p < 0.01), tocopherol (8.46 ± 2.91 vs. 9.79 ± 4.53 g/d,
        p = 0.032), vitamin B 1 (1.91 ± 0.87 vs. 2.3 ± 0.82 g/d, p = 0.01),
        folate (528 ± 0.61 vs. 574.39 ± 95.19 g/d, p = 0.01), biotin
        (26.76 ± 3.75 vs. 29.33 ± 6.61 g/d, p < 0.01) and higher intake of
        calorie (2500.48 ± 165.87 vs. 2594.64 ± 333.4 g/d, p = 0.021), fat
        (86.57 ± 10.38 vs. 93.25, ± 17.13 p < 0.01), fluoride
        (13967.59 ± 5662.25 vs. 11112.32 ± 3051.44 g/d, p < 0.01), vitamin A
        (819.7 ± 251.03 vs. 712.76 ± 113.86 g/d, p = 0.01), and vitamin K
        (157.9 ± 30.4 vs. 146.74 ± 21.64 g/d, p = 0.03).
      - >-
        All concentration estimates are standardized by faecal weight and
        depicted as concentration per gram of faeces. 

        All quantitative PCR reactions were conducted in 12.5 μl volumes using
        the SYBR green Master Mix (Roche). Quantitative PCR experiments were
        conducted on a Lightcycler LC480 instrument (Roche). Template quantity
        and quality was assessed using a Nanodrop spectrophotometer. Abundance
        estimates are standardized to the concentration of input DNA per
        reaction and are represented as copies per nanogram of faecal DNA.
        Template extraction for quantification of faecal bacteria loads: DNA was
        extracted from fresh faecal pellets using the PowerFecal DNA Isolation
        Kit (Mo Bio) following kit instructions. Bacterial loads were quantified
        using previously validated bacterial group-specific 16S primers. 

        Statistics were carried out using JMP9.0 (SAS), Prism 6.0 (Graphpad) and
        R software. permutational analysis of variance was used for hypothesis
        testing of significance between groups shown in PcoA plots.
      - >-
        Postmenopausal diabetic women are at higher risk to develop
        cardiovascular diseases (CVD) compared with nondiabetic women.
        Alterations in cardiac cellular metabolism caused by changes in sirtuins
        are one of the main causes of CVD in postmenopausal diabetic women.
        Several studies have demonstrated the beneficial actions of the G
        protein-coupled estrogen receptor (GPER) in postmenopausal diabetic CVD.
        However, the molecular mechanisms by which GPER has a cardioprotective
        effect are still not well understood. In this study, we used an
        ovariectomized (OVX) type-two diabetic (T2D) rat model induced by
        high-fat diet/streptozotocin to investigate the effect of G-1
        (GPER-agonist) on sirtuins, and their downstream pathways involved in
        regulation of cardiac metabolism and function. Animals were divided into
        five groups: Sham-Control, T2D, OVX+T2D, OVX+T2D+Vehicle, and
        OVX+T2D+G-1. G-1 was administrated for six weeks.
  - source_sentence: >-
      Represent this sentence for searching relevant passages: Why might a VRAM
      flap be a more optimal choice for patients with an end colostomy?
    sentences:
      - >-
        As they will have an end colostomy, which will be their only stoma, then
        a VRAM flap is a more optimal choice given the bulk and ability to fill
        dead space with this flap. Very few patients had infection or dehiscence
        in the postoperative period. Donor-site hernia is a concern with the
        VRAM flap, particularly given an open very large laparotomy incision
        which may often be a reoperation. This occurred in 9.5% of the VRAM
        patients, and the same number of patients required a delayed reoperation
        which was on an elective basis. VRAM, as well as ALT flaps can be used
        to restore the anatomy of the pelvic floor preventing herniation into
        the resection space. The ‘marine patch’ principle applies where the flap
        lies on the side of hydrostatic pressure, so even if there is perineal
        skin breakdown then the muscle flap component still provides cover for
        the abdominal contents. Compared with Baird and colleagues, we reserved
        VRAM flaps for this reason to APR and ELAPE patients. VRAM is not used
        in exenteration in our centre due to two stomas being formed during
        urinary diversion.
      - "In the present study, we used a recently developed novel steatohepatitis-inducing HFD, STHD-01 , to induce NASH. This novel HFD contains a high amount of cholesterol, which is not contained in conventionally used HFDs, and induces the development of severe NASH, while conventionally-used HFDs only induce mild to moderate NASH in a shorter period of time. Another specific feature of STHD-01 is that STHD-01 does not affect fasting blood glucose levels (Additional\_file\_). While certain type of diet, such as methionine- and choline-deficient diet (MCD), can also cause an advanced NASH , this diet decreases fasting blood glucose levels in experimental animals. Since non-overweight human patients with NAFLD do not show decreased fasting blood glucose levels compared to non-fatty liver disease patients , STHD-01 is a better approximation of the clinical condition. One obvious difference in the phenotypes between the mice fed with the STHD-01 and the conventional HFD is body weight gain."
      - >-
        Only 107 (13.8%) were satisfied, and 667 (84%) were dissatisfied.
        Regarding the reasons for dissatisfaction, 355 (45.9%) subjects reported
        that they did not get enough explanation, 292 (37.7%) reported that they
        did not get enough investigations, and only 20 (2.6%) thought that they
        did not get enough medications, as shown in Figure. 

        Of 863 subjects with heartburn, QoL was not affected at all in 295
        (34%), a little in 210 (24%), somewhat in 125 (15%), a lot in 208 (24%),
        and a great deal in 25 (3%) subjects. Considering a lot and a great deal
        as the significant impairment of QoL, 233 (27%) of the subjects had
        impaired QoL due to heartburn. 

        This cross‐sectional study conducted among the adult population in a
        rural community of Bangladesh found that about 26% of the population had
        heartburn, 11% chest pain, 8% globus, and 4% had dysphagia. One‐third of
        the study population had at least one esophageal symptom.
  - source_sentence: >-
      Represent this sentence for searching relevant passages: What percentage
      of the UAE's population resides in Sharjah?
    sentences:
      - >-
        Currently, there is a scarcity of data about the practice and impact of
        OTC medication usage among pregnant women in UAE. Accordingly, this
        study was planned and designed with the aim of exploring the awareness
        and assessing the usage of OTC medications among pregnant women in
        Sharjah, UAE. 

        The study was conducted after the approval of the University of Sharjah
        Ethics Committee, Sharjah, UAE (reference number: REC-16-10-03-01-S). 

        A cross-sectional survey was conducted to assess the level of awareness
        and knowledge of pregnant women concerning OTC drugs. The study took
        place in the Emirate of Sharjah, UAE, over a period of three months
        (October to December 2016). 

        Sharjah is the third largest of the seven emirates that make up the UAE
        and is the only one to have land on both the Arabian Gulf Coast and the
        Gulf of Oman. Residents of Sharjah represent around 19% of the UAE's
        population (4.76 million) (Ministry of Economy, 2008). Within the UAE,
        it has been reported that the crude birth rate or birth rate per 1,000
        population was 15.54 during the year of 2014.
      - >-
        However, following a more painful surgery, children in the VR group
        needed rescue analgesia significantly less often ( p = 0.002). In 2021,
        a total of 50 children aged 6–12-years old were included in a RCT
        evaluating the effect of VR compared to standard screen TV in reducing
        anxiety for buccal infiltration anesthesia. No significant difference
        was observed between the groups, but female and younger patients showed
        higher pain scores during the dentistry procedure. Two recent
        meta-analyses that included a maximum of 17 studies evaluating the
        effect of VR on pain and anxiety in a pediatric population concluded
        that VR is an effective distraction intervention to reduce pain and
        anxiety in children. 

        Finally, other medical fields have also explored the role of VR in
        anxiety reduction. In gastroenterology, VR has been used prior to
        endoscopic procedures to reduce anxiety and has shown promising results,
        reducing anxiety significantly in patients with a higher anxiety level
        (STAI-score ≥ 45) at baseline ( p = 0.007).
      - >-
        Picrosirius Red staining also demonstrated an increase in total collagen
        deposition in the right carotid artery due to TAC-induced vascular
        changes. Alamandine treatment effectively prevented the increase in
        reactive oxygen species production and depletion of nitric oxide levels,
        which were induced by TAC. Finally, alamandine treatment was also shown
        to prevent the increased expression of nuclear factor erythroid
        2-related factor 2 and 3-nitrotyrosine that were induced by TAC. Our
        results suggest that alamandine can effectively attenuate
        pathophysiological stress in the right carotid artery of animals
        subjected to TAC.
  - source_sentence: >-
      Represent this sentence for searching relevant passages: What are some
      effects of maternal iron deficiency on adult male offspring development?
    sentences:
      - >-
        Parents report encouraging their children to engage in “healthy”
        lifestyle choices, including making alterations to diet, physical
        activity (PA), and sleep behavior, which may (1) help parents feel more
        in control over the impact of the condition, and (2) allow them gain a
        more positive outlook on the future. Unfortunately, even in the adult MS
        literature, there is insufficient evidence to make clinical
        recommendations regarding lifestyle modifications. Improving the body of
        literature on modifiable lifestyle factors in pediatric MS with the goal
        of creating guidelines that will help POMS patients and their parents
        deal with these difficult decisions is needed. 

        Our objective in this manuscript is to summarize and identify gaps in
        current research on modifiable lifestyle factors and pediatric MS. Two
        questions guided this review: (1) Which modifiable lifestyle factors
        have been investigated in the context of POMS? And (2) which factors
        have been shown to play a role in the risk of POMS, disease course, or
        quality of life? 

        We used the Arksey and O’Malley framework to guide this review.
      - >-
        The mRNA expression levels of the OMH-treated HT-115 cells indicated
        that the cytosolic CYP1A levels were two-fold upregulated. In addition,
        OMH triggers the mitochondrial release of cytochrome c, which stabilize
        the fundamental oxido-reduction cycle in mitochondria. The activation of
        CYP1A effectively controls the pro-oxidants and oxidative stress in
        colon cancer cells further, suppressing the proinflammatory cytokines
        IL-1β and TNF-α, which favors the deactivation of malignant cell
        apoptosis inhibitor NF-kB in colon cancer cells. The observed
        antioxidant capacity neutralizes proinflammatory TNF-α/IL-1β, inhibiting
        protumorigenic COX-2/PGE-2 and stimulating the apoptosis mechanism via
        the inhibition of NF-kB, an apoptosis inhibitor. OMH effectively
        maintains the balance between Bcl-2 and Bax (Bcl-2-associated X
        pro-apoptotic gene) and inclines the cells to apoptotic stimulation.
      - >-
        We found three differentially abundant taxonomic classes in the IDD
        group using an LDA effect size calculation with an LDA score higher than
        4.0. The results showed that the Bacteroidaceae genus Bacteroides and
        Lachnospiraceae genus Marvinbryantia were significantly increased in
        rats in the IDD group compared to rats in the other groups (C). 

        In this study, we showed that maternal iron deficiency may program and
        alter adult male offspring development with regard to spatial learning
        and memory, dorsal hippocampus BDNF expression, gut microbiota, and SCFA
        concentrations. Our results showed that the adult male offspring of rats
        that were fed a low-iron diet before pregnancy and throughout the
        lactation period had (1) spatial deficits via a Morris water maze
        evaluation; (2) decreased dorsal hippocampal BDNF mRNA and protein
        concentrations accompanied by a low TrkB abundance; (3) a decreased
        plasma acetate concentration without changes in butyrate and propionate
        concentrations; (4) enrichment of the Bacteroidaceae genus Bacteroides
        and Lachnospiraceae genus Marvinbryantia.
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - cosine_accuracy@1
  - cosine_accuracy@3
  - cosine_accuracy@5
  - cosine_accuracy@10
  - cosine_precision@1
  - cosine_precision@3
  - cosine_precision@5
  - cosine_precision@10
  - cosine_recall@1
  - cosine_recall@3
  - cosine_recall@5
  - cosine_recall@10
  - cosine_ndcg@10
  - cosine_mrr@10
  - cosine_map@100
model-index:
  - name: SentenceTransformer based on BAAI/bge-small-en-v1.5
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: Unknown
          type: unknown
        metrics:
          - type: cosine_accuracy@1
            value: 0.5853673532124193
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7196126652320934
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.7634798647402398
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8083922533046418
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.5853673532124193
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.2398708884106978
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.15269597294804796
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.0808392253304642
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.5853673532124193
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7196126652320934
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.7634798647402398
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8083922533046418
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6971481810101028
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6614873816111168
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6662955818767544
            name: Cosine Map@100

SentenceTransformer based on BAAI/bge-small-en-v1.5

This is a sentence-transformers model finetuned from BAAI/bge-small-en-v1.5 on the csv dataset. It maps sentences & paragraphs to a 384-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: BAAI/bge-small-en-v1.5
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 384 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • csv

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, '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("sentence_transformers_model_id")
# Run inference
sentences = [
    'Represent this sentence for searching relevant passages: What are some effects of maternal iron deficiency on adult male offspring development?',
    'We found three differentially abundant taxonomic classes in the IDD group using an LDA effect size calculation with an LDA score higher than 4.0. The results showed that the Bacteroidaceae genus Bacteroides and Lachnospiraceae genus Marvinbryantia were significantly increased in rats in the IDD group compared to rats in the other groups (C). \nIn this study, we showed that maternal iron deficiency may program and alter adult male offspring development with regard to spatial learning and memory, dorsal hippocampus BDNF expression, gut microbiota, and SCFA concentrations. Our results showed that the adult male offspring of rats that were fed a low-iron diet before pregnancy and throughout the lactation period had (1) spatial deficits via a Morris water maze evaluation; (2) decreased dorsal hippocampal BDNF mRNA and protein concentrations accompanied by a low TrkB abundance; (3) a decreased plasma acetate concentration without changes in butyrate and propionate concentrations; (4) enrichment of the Bacteroidaceae genus Bacteroides and Lachnospiraceae genus Marvinbryantia.',
    'Parents report encouraging their children to engage in “healthy” lifestyle choices, including making alterations to diet, physical activity (PA), and sleep behavior, which may (1) help parents feel more in control over the impact of the condition, and (2) allow them gain a more positive outlook on the future. Unfortunately, even in the adult MS literature, there is insufficient evidence to make clinical recommendations regarding lifestyle modifications. Improving the body of literature on modifiable lifestyle factors in pediatric MS with the goal of creating guidelines that will help POMS patients and their parents deal with these difficult decisions is needed. \nOur objective in this manuscript is to summarize and identify gaps in current research on modifiable lifestyle factors and pediatric MS. Two questions guided this review: (1) Which modifiable lifestyle factors have been investigated in the context of POMS? And (2) which factors have been shown to play a role in the risk of POMS, disease course, or quality of life? \nWe used the Arksey and O’Malley framework to guide this review.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Evaluation

Metrics

Information Retrieval

• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.5854
cosine_accuracy@3	0.7196
cosine_accuracy@5	0.7635
cosine_accuracy@10	0.8084
cosine_precision@1	0.5854
cosine_precision@3	0.2399
cosine_precision@5	0.1527
cosine_precision@10	0.0808
cosine_recall@1	0.5854
cosine_recall@3	0.7196
cosine_recall@5	0.7635
cosine_recall@10	0.8084
cosine_ndcg@10	0.6971
cosine_mrr@10	0.6615
cosine_map@100	0.6663

Training Details

Training Dataset

csv

• Dataset: csv
• Size: 650,596 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 16 tokens mean: 26.5 tokens max: 65 tokens	min: 25 tokens mean: 229.67 tokens max: 512 tokens

• Samples:

  anchor	positive
  Represent this sentence for searching relevant passages: What conditions are excluded as secondary causes of hypercholesterolemia?	In addition, no abnormalities were revealed under physical examination. The exclusion criteria comprised secondary causes of hypercholesterolemia, including hypothyroidism, kidney diseases, poorly-controlled diabetes, cholestasis or the use of drugs impairing lipid metabolism. The investigation was approved by the Bioethics Committee of the Medical University of Lodz (RNN/191/21/KE). Informed consent was obtained from all participants. All methods were carried out in accordance with relevant guidelines and regulations. All participants were interviewed for their personal history of diabetes, hypertension, smoking, cardiovascular disease, pharmacological treatment, family history of hypercholesterolemia and cardiovascular disease. During the same visit, a physical examination for the presence of corneal arcus and tendon xanthomas was performed. In both the control and research groups, peripheral blood mononuclear cells (PBMCs) and serum were isolated from peripheral whole blood. All...
  Represent this sentence for searching relevant passages: What type of mannose linkage in side chains has the highest impact on antibody response?	On the other hand, side chains with β-(1→2)-linked mannose residues, which have the highest impact on antibody response , were found only in Candida spp.. The oligomannoside sequence within S. cerevisiae mannan corresponding to antibodies associated with Crohn’s disease was assigned to be the following mannotetraoside: Man(1→3)Man(1→2)Man(1→2)Man , which is illustrated in. Therefore, the corresponding oligosaccharide 1 was selected in this study as a basis for the creation of structurally related glycoarray. Ligands 2 and 3 stem from 1 after formally replacing the terminal α-(1→3)-mannoside fragment with α-(1→2)- and β-(1→2)-mannoside units, respectively. Additional glycosylation of ligand 1 leads to the formation of ligands 4 and 5.
  Represent this sentence for searching relevant passages: How do fluctuations in nest temperature affect bumblebee colonies in aboveground nest boxes?	Impairments to colony function, as a result a sublethal environmental stressors, are linked with reduced colony success , therefore, combined increases in worker abandonment and reduced offspring production may act to have the greatest impact on bumblebee colony success under chronic heat stress. The results obtained from our laboratory study inform about the capacity of bumblebee colonies to cope with chronic warm temperatures, but there are several distinctions when transposed to natural settings. Conditions used correspond more to surface or aboveground nesting that provide minor buffering from the environment. Underground nest sites are the most frequently observed nesting strategies across multiple bumblebee species, including B. impatiens. However, surface or aboveground nest sites combined are almost as frequently reported for natural settings and even more frequent when nesting in artificial nest such as human made structures. Aboveground temperatures can cause wide fluctuatio...

• Loss: CachedGISTEmbedLoss with these parameters:

  {'guide': SentenceTransformer(
    (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: NewModel 
    (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
    (2): Dense({'in_features': 1024, 'out_features': 1024, 'bias': True, 'activation_function': 'torch.nn.modules.linear.Identity'})
  ), 'temperature': 0.01}
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 32768
• num_train_epochs: 8
• lr_scheduler_type: cosine
• warmup_ratio: 0.1
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 32768
• per_device_eval_batch_size: 8
• 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: 5e-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: 8
• max_steps: -1
• lr_scheduler_type: cosine
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.1
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• save_safetensors: True
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• no_cuda: False
• use_cpu: False
• use_mps_device: False
• seed: 42
• data_seed: None
• jit_mode_eval: False
• use_ipex: False
• bf16: False
• fp16: False
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: 0
• ddp_backend: None
• tpu_num_cores: None
• tpu_metrics_debug: False
• debug: []
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• past_index: -1
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: []
• fsdp_min_num_params: 0
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• fsdp_transformer_layer_cls_to_wrap: None
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch
• optim_args: None
• adafactor: False
• group_by_length: False
• length_column_name: length
• ddp_find_unused_parameters: None
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• dataloader_pin_memory: True
• dataloader_persistent_workers: False
• skip_memory_metrics: True
• use_legacy_prediction_loop: False
• push_to_hub: False
• resume_from_checkpoint: None
• hub_model_id: None
• hub_strategy: every_save
• hub_private_repo: None
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• include_for_metrics: []
• eval_do_concat_batches: True
• fp16_backend: auto
• push_to_hub_model_id: None
• push_to_hub_organization: None
• mp_parameters:
• auto_find_batch_size: False
• full_determinism: False
• torchdynamo: None
• ray_scope: last
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• dispatch_batches: None
• split_batches: None
• include_tokens_per_second: False
• include_num_input_tokens_seen: False
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• eval_on_start: False
• use_liger_kernel: False
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Click to expand

Epoch	Step	Training Loss	cosine_ndcg@10
0.0526	1	7.2666	-
0.1053	2	7.2688	-
0.1579	3	6.8798	-
0.2105	4	6.0896	-
0.2632	5	5.1499	0.5392
0.3158	6	4.2179	-
0.3684	7	3.4166	-
0.4211	8	2.9593	-
0.4737	9	2.8846	-
0.5263	10	2.8879	0.5541
0.5789	11	2.728	-
0.6316	12	2.5792	-
0.6842	13	2.4242	-
0.7368	14	2.2856	-
0.7895	15	2.2488	0.5852
0.8421	16	2.1646	-
0.8947	17	2.0432	-
0.9474	18	1.9749	-
1.0	19	1.8132	-
1.0526	20	1.8851	0.6135
1.1053	21	1.8024	-
1.1579	22	1.777	-
1.2105	23	1.7047	-
1.2632	24	1.6751	-
1.3158	25	1.6875	0.6283
1.3684	26	1.6396	-
1.4211	27	1.5756	-
1.4737	28	1.5591	-
1.5263	29	1.533	-
1.5789	30	1.5035	0.6449
1.6316	31	1.4705	-
1.6842	32	1.4446	-
1.7368	33	1.4092	-
1.7895	34	1.4139	-
1.8421	35	1.3996	0.6557
1.8947	36	1.365	-
1.9474	37	1.3397	-
2.0	38	1.2443	-
2.0526	39	1.3322	-
2.1053	40	1.2862	0.6632
2.1579	41	1.2965	-
2.2105	42	1.2544	-
2.2632	43	1.2474	-
2.3158	44	1.2748	-
2.3684	45	1.2509	0.6688
2.4211	46	1.2271	-
2.4737	47	1.2172	-
2.5263	48	1.2263	-
2.5789	49	1.1919	-
2.6316	50	1.1962	0.6748
2.6842	51	1.1732	-
2.7368	52	1.1683	-
2.7895	53	1.1711	-
2.8421	54	1.1783	-
2.8947	55	1.1353	0.6784
2.9474	56	1.1301	-
3.0	57	1.0551	-
3.0526	58	1.1436	-
3.1053	59	1.0967	-
3.1579	60	1.1259	0.6822
3.2105	61	1.085	-
3.2632	62	1.1107	-
3.3158	63	1.104	-
3.3684	64	1.1113	-
3.4211	65	1.0884	0.6849
3.4737	66	1.079	-
3.5263	67	1.0946	-
3.5789	68	1.0751	-
3.6316	69	1.0585	-
3.6842	70	1.0601	0.6877
3.7368	71	1.0576	-
3.7895	72	1.0558	-
3.8421	73	1.0642	-
3.8947	74	1.0349	-
3.9474	75	1.0368	0.6889
4.0	76	0.9558	-
4.0526	77	1.0487	-
4.1053	78	1.0164	-
4.1579	79	1.0359	-
4.2105	80	1.0095	0.6908
4.2632	81	1.0194	-
4.3158	82	1.0359	-
4.3684	83	1.0266	-
4.4211	84	1.0161	-
4.4737	85	1.0188	0.6913
4.5263	86	1.0265	-
4.5789	87	1.0193	-
4.6316	88	1.0052	-
4.6842	89	0.9994	-
4.7368	90	1.0024	0.6934
4.7895	91	1.0134	-
4.8421	92	1.0259	-
4.8947	93	0.9807	-
4.9474	94	0.9947	-
5.0	95	0.9139	0.6945
5.0526	96	0.9956	-
5.1053	97	0.9615	-
5.1579	98	0.9942	-
5.2105	99	0.9616	-
5.2632	100	0.9848	0.6947
5.3158	101	0.9967	-
5.3684	102	0.9861	-
5.4211	103	0.9694	-
5.4737	104	0.984	-
5.5263	105	0.9953	0.6953
5.5789	106	0.987	-
5.6316	107	0.9745	-
5.6842	108	0.9582	-
5.7368	109	0.957	-
5.7895	110	0.9826	0.6960
5.8421	111	0.9911	-
5.8947	112	0.96	-
5.9474	113	0.9593	-
6.0	114	0.8886	-
6.0526	115	0.9722	0.6963
6.1053	116	0.9507	-
6.1579	117	0.9767	-
6.2105	118	0.9394	-
6.2632	119	0.9569	-
6.3158	120	0.9674	0.6965
6.3684	121	0.9674	-
6.4211	122	0.9606	-
6.4737	123	0.96	-
6.5263	124	0.9767	-
6.5789	125	0.9664	0.6968
6.6316	126	0.948	-
6.6842	127	0.9581	-
6.7368	128	0.9491	-
6.7895	129	0.9627	-
6.8421	130	0.9723	0.6971
6.8947	131	0.9447	-
6.9474	132	0.9502	-
7.0	133	0.8796	-
7.0526	134	0.9589	-
7.1053	135	0.9377	0.6971
7.1579	136	0.9573	-
7.2105	137	0.9369	-
7.2632	138	0.9559	-
7.3158	139	0.9662	-
7.3684	140	0.9615	0.6971
7.4211	141	0.9555	-
7.4737	142	0.9579	-
7.5263	143	0.9719	-
7.5789	144	0.9664	-
7.6316	145	0.9554	0.6972
7.6842	146	0.9526	-
7.7368	147	0.9456	-
7.7895	148	0.9621	-
7.8421	149	0.9669	-
7.8947	150	0.9473	0.6971
7.9474	151	0.9519	-
8.0	152	0.8705	-

Framework Versions

• Python: 3.11.9
• Sentence Transformers: 3.3.1
• Transformers: 4.47.0
• PyTorch: 2.5.1+cu124
• Accelerate: 1.2.0
• Datasets: 2.19.2
• Tokenizers: 0.21.0

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}