metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:35934
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
widget:
  - source_sentence: >-
      Stating purpose of the current research with reference to gaps or issues
      in the literature
    sentences:
      - >-
        During the 15-year study, 10% of the osseointegrated implants in the
        edentulous jaw showed signs of peri-implantitis, leading to their
        failure.
      - >-
        This paper provides an in-depth exploration of the qualitative case
        study methodology, addressing the lack of comprehensive guidance for
        novice researchers in this area.
      - >-
        As a novice researcher in management science, I have been drawn to the
        qualitative case study methodology due to its ability to provide rich,
        in-depth insights into complex real-world situations.
  - source_sentence: Indicating missing, weak, or contradictory evidence
    sentences:
      - >-
        This paper contributes to the literature on the financial system by
        examining the relationship between bank size, bank capital, and the bank
        lending channel using a unique dataset of US banks during the global
        financial crisis.
      - >-
        A total of 150 patients with a clinical diagnosis of osteoarthritis of
        the hip or knee, according to the American College of Rheumatology
        criteria, were included in the study.
      - >-
        Despite the widespread use of the WOMAC (Western Ontario and McMaster
        Universities Osteoarthritis Index) questionnaire in clinical practice
        and research, there is a lack of consensus regarding its responsiveness
        to antirheumatic drug therapy in patients with osteoarthritis of the hip
        or knee.
  - source_sentence: >-
      Establishing the importance of the topic for the world or society: time
      frame given
    sentences:
      - >-
        The Th/Hf ratios of the basaltic lavas from the British Tertiary
        Volcanic Province range from 4.2 to 5.5, as shown in Table 1.
      - >-
        The use of organometal halide perovskites as visible-light sensitizers
        for photovoltaic cells has gained significant attention in the
        optoelectronics community due to their promising photovoltaic
        performance and cost-effective fabrication since the late 2000s.
      - >-
        Table 1 summarizes the power conversion efficiencies (PCEs) and
        certifications of the best-performing perovskite solar cells reported in
        the literature.
  - source_sentence: Describing the research design and the methods used
    sentences:
      - >-
        This study aims to evaluate the efficacy and safety of preoperative
        radiotherapy followed by total mesorectal excision in the treatment of
        resectable rectal cancer.
      - >-
        TREE-PUZZLE's parallel computing implementation significantly reduces
        the time required for maximum likelihood phylogenetic analysis compared
        to traditional methods, supporting previous findings of the importance
        of parallelization in phylogenetics.
      - >-
        This study investigates the efficacy of preoperative radiotherapy
        followed by total mesorectal excision in the treatment of resectable
        rectal cancer.
  - source_sentence: 'Surveys and interviews: Introducing excerpts from interview data'
    sentences:
      - >-
        Previous research on international trade under the WTO regime has
        explored various approaches to understanding the uneven promotion of
        trade (Hoekstra & Kostecki, 2001; Cline, 2004, ...).
      - >-
        Through surveys and interviews, multiliterate teachers expressed a
        shared belief in the importance of fostering students' ability to
        navigate multiple discourse communities.
      - >-
        The authors employ a constructivist approach to learning, where students
        build knowledge through active engagement with multimedia texts and
        collaborative discussions.
datasets:
  - Corran/SciGenTriplets
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
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: SciGen Eval Set
          type: SciGen-Eval-Set
        metrics:
          - type: cosine_accuracy@1
            value: 0.8918076580587712
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.9307658058771149
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.9481300089047195
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.9668299198575245
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.8918076580587712
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.3102552686257049
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.18962600178094388
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.09668299198575243
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.8918076580587712
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.9307658058771149
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.9481300089047195
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.9668299198575245
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.9279217256301748
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.9156546382281018
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.9171082586239344
            name: Cosine Map@100

SentenceTransformer

This is a sentence-transformers model trained on the sci_gen_colbert_triplets dataset. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

Model Type: Sentence Transformer
Maximum Sequence Length: inf tokens
Output Dimensionality: 768 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- sci_gen_colbert_triplets

Model Sources

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

Full Model Architecture

SentenceTransformer(
  (0): StaticEmbedding(
    (embedding): EmbeddingBag(30522, 768, mode='mean')
  )
)

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("Corran/SciGenNomicEmbedStatic")
# Run inference
sentences = [
    'Surveys and interviews: Introducing excerpts from interview data',
    "Through surveys and interviews, multiliterate teachers expressed a shared belief in the importance of fostering students' ability to navigate multiple discourse communities.",
    'The authors employ a constructivist approach to learning, where students build knowledge through active engagement with multimedia texts and collaborative discussions.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

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

Evaluation

Metrics

Information Retrieval

Dataset: SciGen-Eval-Set
Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.8918
cosine_accuracy@3	0.9308
cosine_accuracy@5	0.9481
cosine_accuracy@10	0.9668
cosine_precision@1	0.8918
cosine_precision@3	0.3103
cosine_precision@5	0.1896
cosine_precision@10	0.0967
cosine_recall@1	0.8918
cosine_recall@3	0.9308
cosine_recall@5	0.9481
cosine_recall@10	0.9668
cosine_ndcg@10	0.9279
cosine_mrr@10	0.9157
cosine_map@100	0.9171

Training Details

Training Dataset

sci_gen_colbert_triplets

Dataset: sci_gen_colbert_triplets at 44071bd
Size: 35,934 training samples
Columns: query, positive, and negative

Approximate statistics based on the first 1000 samples:

	query	positive	negative
type	string	string	string
details	min: 20 characters mean: 50.28 characters max: 120 characters	min: 0 characters mean: 206.53 characters max: 401 characters	min: 96 characters mean: 209.67 characters max: 418 characters

Samples:

query	positive	negative
`Previous research: highlighting negative outcomes`	`Despite the widespread use of seniority-based wage systems in labor contracts, previous research has highlighted their negative outcomes, such as inefficiencies and demotivating effects on workers.`	`This paper, published in 1974, was among the first to establish the importance of rank-order tournaments as optimal labor contracts in microeconomics.`
`Synthesising sources: contrasting evidence or ideas`	`Despite the observed chronic enterocolitis in Interleukin-10-deficient mice, some studies suggest that this cytokine plays a protective role in intestinal inflammation in humans (Kurimoto et al., 2001).`	`Chronic enterocolitis developed in Interleukin-10-deficient mice, characterized by inflammatory cell infiltration, epithelial damage, and increased production of pro-inflammatory cytokines.`
`Previous research: Approaches taken`	`Previous research on measuring patient-relevant outcomes in osteoarthritis has primarily relied on self-reported measures, such as the Western Ontario and McMaster Universities Arthritis Index (WOMAC) (Bellamy et al., 1988).`	`The WOMAC (Western Ontario and McMaster Universities Osteoarthritis Index) questionnaire has been widely used in physical therapy research to assess the impact of antirheumatic drug therapy on patient-reported outcomes in individuals with hip or knee osteoarthritis.`

Loss: MatryoshkaLoss with these parameters:

{
    "loss": "MultipleNegativesRankingLoss",
    "matryoshka_dims": [
        768,
        384,
        256,
        128,
        64,
        32
    ],
    "matryoshka_weights": [
        1,
        1,
        1,
        1,
        1,
        1
    ],
    "n_dims_per_step": -1
}

Evaluation Dataset

sci_gen_colbert_triplets

Dataset: sci_gen_colbert_triplets at 44071bd
Size: 4,492 evaluation samples
Columns: query, positive, and negative

Approximate statistics based on the first 1000 samples:

	query	positive	negative
type	string	string	string
details	min: 20 characters mean: 50.59 characters max: 120 characters	min: 98 characters mean: 203.98 characters max: 448 characters	min: 36 characters mean: 204.82 characters max: 422 characters

Samples:

query	positive	negative
`Providing background information: reference to the purpose of the study`	`This study aimed to investigate the impact of socioeconomic status on child development, specifically focusing on cognitive, language, and social-emotional domains.`	`Children from high socioeconomic status families showed significantly higher IQ scores (M = 112.5, SD = 5.6) compared to children from low socioeconomic status families (M = 104.3, SD = 6.2) in the verbal IQ subtest.`
`Providing background information: reference to the literature`	`According to previous studies using WinGX suite for small-molecule single-crystal crystallography, the optimization of crystal structures leads to improved accuracy in determining atomic coordinates.`	`This paper describes the WinGX suite, a powerful tool for small-molecule single-crystal crystallography that significantly advances the field of crystallography by streamlining data collection and analysis.`
`General comments on the relevant literature`	`Polymer brushes have gained significant attention in the field of polymer science due to their unique properties, such as controlled thickness, high surface density, and tunable interfacial properties.`	`Despite previous reports suggesting that polymer brushes with short grafting densities exhibit poorer performance in terms of adhesion and stability compared to those with higher grafting densities (Liu et al., 2010), our results indicate that the opposite is true for certain types of polymer brushes.`

Loss: MatryoshkaLoss with these parameters:

{
    "loss": "MultipleNegativesRankingLoss",
    "matryoshka_dims": [
        768,
        384,
        256,
        128,
        64,
        32
    ],
    "matryoshka_weights": [
        1,
        1,
        1,
        1,
        1,
        1
    ],
    "n_dims_per_step": -1
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 4096
per_device_eval_batch_size: 4096
learning_rate: 0.02
num_train_epochs: 50
warmup_ratio: 0.1
fp16: True

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 4096
per_device_eval_batch_size: 4096
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: 0.02
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 50
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.1
warmup_steps: 0
log_level: passive
log_level_replica: warning
log_on_each_node: True
logging_nan_inf_filter: True
save_safetensors: True
save_on_each_node: False
save_only_model: False
restore_callback_states_from_checkpoint: False
no_cuda: False
use_cpu: False
use_mps_device: False
seed: 42
data_seed: None
jit_mode_eval: False
use_ipex: False
bf16: False
fp16: 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: batch_sampler
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	Validation Loss	SciGen-Eval-Set_cosine_ndcg@10
-1	-1	-	-	0.0860
1.1111	10	64.4072	61.6146	0.0919
2.2222	20	60.2737	56.0852	0.1130
3.3333	30	53.8742	50.1738	0.1611
4.4444	40	47.9741	45.6099	0.2666
5.5556	50	43.3533	42.3335	0.4579
6.6667	60	39.8746	40.0990	0.6244
7.7778	70	37.4077	38.4205	0.7223
8.8889	80	35.3558	37.0939	0.7847
10.0	90	33.5816	36.0200	0.8248
11.1111	100	32.4019	35.1148	0.8469
12.2222	110	31.3427	34.3602	0.8658
13.3333	120	30.4578	33.7324	0.8788
14.4444	130	29.7019	33.2120	0.8882
15.5556	140	29.1315	32.7679	0.8963
16.6667	150	28.6226	32.3942	0.9016
17.7778	160	28.195	32.0693	0.9061
18.8889	170	27.8242	31.7708	0.9096
20.0	180	27.373	31.5369	0.9137
21.1111	190	27.2436	31.3331	0.9168
22.2222	200	27.0084	31.1571	0.9188
23.3333	210	26.8023	31.0074	0.9205
24.4444	220	26.6754	30.8726	0.9217
25.5556	230	26.4875	30.7545	0.9224
26.6667	240	26.3846	30.6494	0.9236
27.7778	250	26.2546	30.5660	0.9243
28.8889	260	26.1752	30.4826	0.9248
30.0	270	25.9247	30.4060	0.9252
31.1111	280	25.9807	30.3540	0.9261
32.2222	290	25.9153	30.3040	0.9262
33.3333	300	25.8643	30.2585	0.9265
34.4444	310	25.7946	30.2183	0.9270
35.5556	320	25.7723	30.1799	0.9272
36.6667	330	25.7091	30.1539	0.9275
37.7778	340	25.6655	30.1296	0.9275
38.8889	350	25.6465	30.1120	0.9276
40.0	360	25.4654	30.0834	0.9279

Framework Versions

Python: 3.11.11
Sentence Transformers: 3.4.0
Transformers: 4.47.1
PyTorch: 2.5.1+cu121
Accelerate: 1.2.1
Datasets: 3.2.0
Tokenizers: 0.21.0

Citation

BibTeX

Sentence Transformers

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

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning},
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

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}
}