checkthat_task4b__9d3e67ce9794_dev_titles_abstracts_gtr-t5-base_fine_tune_TrueK10

SentenceTransformer based on sentence-transformers/gtr-t5-base

This is a sentence-transformers model finetuned from sentence-transformers/gtr-t5-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: sentence-transformers/gtr-t5-base <!-- at revision 9801579ce813fb37541e6098dffd17959fffcd6e -->
• 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
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: T5EncoderModel 
  (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): Dense({'in_features': 768, 'out_features': 768, 'bias': False, 'activation_function': 'torch.nn.modules.linear.Identity'})
  (3): 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("enridaga/checkthat_task4b__9d3e67ce9794_dev_titles_abstracts_gtr-t5-base_fine_tune_TrueK10")
# Run inference
sentences = [
    'The first person with covid-19 in France was as early as December 2019, not January. "[t]he absence of a connection with china and the lack of recent travel suggest that the disease was already spreading among the French population at the end of December, 2019."',
    'SARS-COV-2 was already spreading in France in late December 2019 Abstract The COVID-19 epidemic is believed to have started in late January 2020 in France. We report here a case of a patient hospitalized in December 2019 in our intensive care, of our hospital in the north of Paris, for hemoptysis with no etiological diagnosis and for which RT-PCR was performed retrospectively on the stored respiratory sample which confirmed the diagnosis of COVID-19 infection. Based on this result, it appears that the COVID-19 epidemic started much earlier.',
    'The History of Methylprednisolone, Ascorbic Acid, Thiamine, and Heparin Protocol and I-MASK+ Ivermectin Protocol for COVID-19 An alliance of established experts on critical care, Front Line COVID-19 Critical Care Alliance (FLCCC), has published two protocols for treatment of COVID-19. The first one, methylprednisolone, ascorbic acid, thiamine, and heparin (MATH+), is intended for hospital and intensive care unit treatment of pulmonary phases of the disease. It is based on affordable, commonly available components: anti-inflammatory corticosteroids (methylprednisolone, "M"), high-dose vitamin C infusion (ascorbic acid, "A"), vitamin B1 (thiamine, "T"), anticoagulant heparin ("H"), antiparasitic agent ivermectin, and supplemental components ("+") including melatonin, vitamin D, elemental zinc, and magnesium. The MATH+ protocol has received scarce attention due to the World Health Organization (WHO) advising against the use of corticosteroids in the beginning of the pandemic. In addition, randomized controlled clinical trials were required as a condition for adoption of the protocol. As the hospital mortality rate of MATH+ treated patients was approximately a quarter of the rate of patients receiving a standard of care, the authors of the protocol considered performing such trials unethical. Other parties have later performed clinical trials with corticosteroids and anticoagulants, which has led to a more widespread adoption of these components. In October 2020, ivermectin was upgraded from an optional component to an essential component of the protocol. According to the authors, ivermectin is considered the first agent effective for both prophylaxis (prevention) of COVID-19 and for treatment of all phases of COVID-19 including outpatient treatment of the early symptomatic phase. Therefore, at the end of October 2020, a separate ivermectin-based I-MASK+ protocol for prophylaxis and early outpatient treatment of COVID-19 was published.',
]
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.* --> <!-- ## 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: 1,119 training samples
• Columns: <code>sentence_0</code>, <code>sentence_1</code>, and <code>sentence_2</code>
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1	sentence_2
  type	string	string	string
  details	<ul><li>min: 9 tokens</li><li>mean: 49.02 tokens</li><li>max: 121 tokens</li></ul>	<ul><li>min: 26 tokens</li><li>mean: 383.13 tokens</li><li>max: 512 tokens</li></ul>	<ul><li>min: 1 tokens</li><li>mean: 384.12 tokens</li><li>max: 512 tokens</li></ul>

• Samples:

  sentence_0	sentence_1	sentence_2
  <code>This meta-analysis evaluated cardiovascular (cv) and kidney impacts of sglt2 inhibitors in patients with type 2 diabetes. risk of major adverse cv events, hospitalisation for heart failure, and kidney disease progression was lowered</code>	<code>Association of SGLT2 Inhibitors With Cardiovascular and Kidney Outcomes in Patients With Type 2 Diabetes: A Meta-analysis. Importance Sodium-glucose cotransporter 2 (SGLT2) inhibitors favorably affect cardiovascular (CV) and kidney outcomes; however, the consistency of outcomes across the class remains uncertain. Objective To perform meta-analyses that assess the CV and kidney outcomes of all 4 available SGLT2 inhibitors in patients with type 2 diabetes. Data Sources A systematic literature search was conducted in PubMed from January 1, 2015, to January 31, 2020. Study Selection One hundred forty-five records were initially identified; 137 were excluded because of study design or topic of interest. As a result, a total of 6 randomized, placebo-controlled CV and kidney outcomes trials of SGLT2 inhibitors in patients with type 2 diabetes were identified, with contributory data from 9 publications. All analyses were conducted on the total patient population of these trials. Data Extractio...</code>	<code>Association of State-Issued Mask Mandates and Allowing On-Premises Restaurant Dining with County-Level COVID-19 Case and Death Growth Rates — United States, March 1–December 31, 2020 CDC recommends a combination of evidence-based strategies to reduce transmission of SARS-CoV-2, the virus that causes COVID-19 (1). Because the virus is transmitted predominantly by inhaling respiratory droplets from infected persons, universal mask use can help reduce transmission (1). Starting in April, 39 states and the District of Columbia (DC) issued mask mandates in 2020. Reducing person-to-person interactions by avoiding nonessential shared spaces, such as restaurants, where interactions are typically unmasked and physical distancing (≥6 ft) is difficult to maintain, can also decrease transmission (2). In March and April 2020, 49 states and DC prohibited any on-premises dining at restaurants, but by mid-June, all states and DC had lifted these restrictions. To examine the association of state-issued...</code>
  <code>It's astonishing, disturbing, incredible that longcovid can cause symptoms across so many organ systems while not resulting in biochemical abnormalities in any organ system! I'm unsure if the nih's 1.4 billion dollars will suffice, it might take 10x as much for answers!</code>	<code>A Longitudinal Study of COVID-19 Sequelae and Immunity: Baseline Findings. BACKGROUND A substantial proportion of persons who develop COVID-19 report persistent symptoms after acute illness. Various pathophysiologic mechanisms have been implicated in the pathogenesis of postacute sequelae of SARS-CoV-2 infection (PASC). OBJECTIVE To characterize medical sequelae and persistent symptoms after recovery from COVID-19 in a cohort of disease survivors and controls. DESIGN Cohort study. (ClinicalTrials.gov: NCT04411147). SETTING National Institutes of Health Clinical Center, Bethesda, Maryland. PARTICIPANTS Self-referred adults with laboratory-documented SARS-CoV-2 infection who were at least 6 weeks from symptom onset were enrolled regardless of presence of PASC. A control group comprised persons with no history of COVID-19 or serologic evidence of SARS-CoV-2 infection, recruited regardless of their current health status. Both groups were enrolled over the same period and from the same geog...</code>	<code>Hospitalizations Associated with COVID-19 Among Children and Adolescents — COVID-NET, 14 States, March 1, 2020–August 14, 2021 Although COVID-19-associated hospitalizations and deaths have occurred more frequently in adults,† COVID-19 can also lead to severe outcomes in children and adolescents (1,2). Schools are opening for in-person learning, and many prekindergarten children are returning to early care and education programs during a time when the number of COVID-19 cases caused by the highly transmissible B.1.617.2 (Delta) variant of SARS-CoV-2, the virus that causes COVID-19, is increasing.§ Therefore, it is important to monitor indicators of severe COVID-19 among children and adolescents. This analysis uses Coronavirus Disease 2019-Associated Hospitalization Surveillance Network (COVID-NET)¶ data to describe COVID-19-associated hospitalizations among U.S. children and adolescents aged 0-17 years. During March 1, 2020-August 14, 2021, the cumulative incidence of COVID-19-associate...</code>
  <code>A recent study conducted by researchers reveals high rates of COVID-19 hospitalizations & readmissions among individuals facing homelessness or imprisonment in the U.S. Some observations... (1/)</code>	<code>Hospitalizations for COVID-19 Among US People Experiencing Incarceration or Homelessness IMPORTANCE: People experiencing incarceration (PEI) and people experiencing homelessness (PEH) have an increased risk of COVID-19 exposure from congregate living, but data on their hospitalization course compared with that of the general population are limited. OBJECTIVE: To compare COVID-19 hospitalizations for PEI and PEH with hospitalizations among the general population. DESIGN, SETTING, AND PARTICIPANTS: This cross-sectional analysis used data from the Premier Healthcare Database on 3415 PEI and 9434 PEH who were evaluated in the emergency department or were hospitalized in more than 800 US hospitals for COVID-19 from April 1, 2020, to June 30, 2021. EXPOSURES: Incarceration or homelessness. MAIN OUTCOMES AND MEASURES: Hospitalization proportions were calculated. and outcomes (intensive care unit admission, invasive mechanical ventilation [IMV], mortality, length of stay, and readmissions) amo...</code>	<code>A Trial of Lopinavir–Ritonavir in Adults Hospitalized with Severe Covid-19 BACKGROUND: No therapeutics have yet been proven effective for the treatment of severe illness caused by SARS-CoV-2. METHODS: We conducted a randomized, controlled, open-label trial involving hospitalized adult patients with confirmed SARS-CoV-2 infection, which causes the respiratory illness Covid-19, and an oxygen saturation (Sao(2)) of 94% or less while they were breathing ambient air or a ratio of the partial pressure of oxygen (Pao(2)) to the fraction of inspired oxygen (Fio(2)) of less than 300 mm Hg. Patients were randomly assigned in a 1:1 ratio to receive either lopinavir–ritonavir (400 mg and 100 mg, respectively) twice a day for 14 days, in addition to standard care, or standard care alone. The primary end point was the time to clinical improvement, defined as the time from randomization to either an improvement of two points on a seven-category ordinal scale or discharge from the hospital, whichever ...</code>

• Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• num_train_epochs: 10
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

<details><summary>Click to expand</summary>

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: no
• prediction_loss_only: True
• per_device_train_batch_size: 8
• 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
• num_train_epochs: 10
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.0
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• save_safetensors: True
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• no_cuda: False
• use_cpu: False
• use_mps_device: False
• seed: 42
• data_seed: None
• jit_mode_eval: False
• use_ipex: False
• bf16: False
• fp16: 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}
• tp_size: 0
• 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
• 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: batch_sampler
• multi_dataset_batch_sampler: round_robin

</details>

Training Logs

Framework Versions

• Python: 3.11.12
• Sentence Transformers: 3.4.1
• Transformers: 4.51.3
• PyTorch: 2.6.0+cu124
• Accelerate: 1.6.0
• Datasets: 3.6.0
• Tokenizers: 0.21.1

Citation

BibTeX

Sentence Transformers

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

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

<!-- ## 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.* -->