Hugging Face's logo

MugheesAwan11
/
bge-base-securiti-dataset-1-v13

Sentence Similarity
sentence-transformers
Safetensors
English
bert
feature-extraction
Generated from Trainer
dataset_size:872
loss:MatryoshkaLoss
loss:MultipleNegativesRankingLoss
Eval Results
text-embeddings-inference
Inference Endpoints
Model card Files Community
bge-base-securiti-dataset-1-v13 / README.md
MugheesAwan11's picture
MugheesAwan11
Add new SentenceTransformer model.
454badc verified
preview code
|
raw
history blame contribute delete
No virus
37.8 kB
metadata
language:
  - en
license: apache-2.0
library_name: sentence-transformers
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:872
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
base_model: BAAI/bge-base-en-v1.5
datasets: []
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
widget:
  - source_sentence: >-
      amendements to PIPA came into force on 05 Auguest 2020. 2 Some parts of
      PIPA also apply to online service providers. 3 The latest amendment to
      PIPA has introduced the concept of ‘pseudonymised data’ for the
      feasibility of data economy. 4 Under the PIPA, all data handlers must
      appoint a chief privacy officer. 5 Cookies, IP information, etc. are also
      regulated by the PIPA as personal information. 6 Breach of a corrective
      order issued by the PIPC can lead to an administrative fine of not more
      than KRW 30 million. ### Forrester Names Securiti a Leader in the Privacy
      Management Wave Q4, 2021 Read the Report ### Securiti named a Leader in
      the IDC MarketScape for Data Privacy Compliance Software Read the Report
      At
    sentences:
      - What recognition did Securiti receive in the field of data privacy?
      - >-
        How does the Office of the Privacy Commissioner educate agencies and
        organisations in breach of the law?
      - >-
        What is the concept of 'pseudonymised data' introduced by the latest
        amendment to PIPA?
  - source_sentence: >-
      18th, 2020, and it has been in effect since then. ## Influence of GDPR It
      is well known that the LGPD was drafted and based on the GDPR, so much so
      that some people call it Brazil’s GDPR. The LGPD contains 65 articles that
      provide individuals with data subject rights, impose obligations upon
      organizations for lawful processing of personal data, require notification
      of data breaches to the supervisory authority and affected data subjects,
      create a national supervisory authority to interpret and enforce the law,
      regulate international transfer of data, define lawful consent collection
      guidelines and impose heavy penalties on violators similar to the GDPR. ##
      Essence of the LGPD Law LGPD provides: 9 data subject rights requests
      exercisable by individual data subjects; 10 legal bases for lawful
      processing; Obligatory and transparent disclosure requirements for
      organizations to contain within their privacy policy; Consent collection
      and management requirements for organizations;
    sentences:
      - >-
        What are the penalties for misusing personal data and obstructing
        investigations under the PDPA and its amendments?
      - >-
        Which data privacy regulation, similar to the GDPR, had a significant
        impact in the US after the promulgation of the GDPR in the EU?
      - >-
        What are the requirements for consent collection and management under
        the LGPD law?
  - source_sentence: >-
      to the Privacy Act of 2020. ## Obligations for Organisations Under the
      Privacy Act 2020 Under the Privacy Act’s jurisdiction, all organizations
      have specific responsibilities or obligations towards their users. The
      most important of these obligations include the following: ### 1\. Lawful
      Purpose Requirements While data processing has become immensely important
      for nearly all businesses, the Privacy Act ensures that such data
      processing can only occur if the organization collecting the data has a
      lawful purpose for the collection and that collection of the information
      is necessary for that purpose. It is also expected that the information
      will be collected directly from the individual concerned. When collecting
      personal information, organizations are required to ensure the individual
      is aware of: The fact that the information is being collected; The purpose
      for which it is being collected; The intended recipients of the
      information; The details of the organization that will be collecting and
      holding the information; Any laws that authorize or
    sentences:
      - >-
        What are the obligations of organizations towards users under the
        Privacy Act of 2020, including lawful purpose and consent requirements?
      - >-
        What is the role of the Spanish Data Protection Agency in enforcing data
        protection legislation in Spain and how does it ensure its effectiveness
        in enforcing the law across the country?
      - >-
        What is the purpose of Kuwait's Data Privacy Protection Regulation
        (DPPR)?
  - source_sentence: >-
      ## Right of Access to Personal Data: What To Know The wealth of data
      available to organizations globally has brought tremendous improvements in
      their ability to target and cater to their customers' needs.
      Organizations... View More September 13, 2023 ## Kuwait's DPPR Kuwait
      didn’t have any data protection law until the Communication and
      Information Technology Regulatory Authority (CITRA) introduced the Data
      Privacy Protection Regulation (DPPR). The... ## Take a Product Tour See
      how easy it is to manage privacy compliance with robotic automation. Watch
      a demo At Securiti, our mission is to enable enterprises to safely harness
      the incredible power of data and the cloud by controlling the complex
      security, privacy and compliance risks. Copyright (C) 2023 Securiti
      Sitemap XML Sitemap #### Newsletter #### Company About Us ,  Personal
      Data: What To Know The wealth of data available to organizations globally
      has brought tremendous improvements in their ability to target and cater
      to their customers' needs. Organizations... View More September 13, 2023
      ## Kuwait's DPPR Kuwait didn’t have any data protection law until the
      Communication and Information Technology Regulatory Authority (CITRA)
      introduced the Data Privacy Protection Regulation (DPPR). The... ## Take a
      Product Tour See how easy it is to manage privacy compliance with robotic
      automation. Watch a demo At Securiti, our mission is to enable enterprises
      to safely harness the incredible power of data and the cloud by
      controlling the complex security, privacy and compliance risks. Copyright
      (C) 2023 Securiti Sitemap XML Sitemap #### Newsletter #### Company About
      Us Careers Contact Us
    sentences:
      - What is the definition of personal data according to the PDPO?
      - >-
        What are the requirements for organizations to notify the regulatory
        authority in case of a data breach according to the PDPL and
        accompanying Regulations?
      - Why did CITRA introduce Kuwait's DPPR?
  - source_sentence: >-
      View Salesforce View Workday View GCP View Azure View Oracle View Learn
      more Regulations Automate compliance with global privacy regulations. US
      California CCPA View US California CPRA View European Union GDPR View
      Thailand’s PDPA View China PIPL View Canada PIPEDA View Brazil's LGPD View
      \+ More View Learn more Roles Identify data risk and enable protection &
      control. Privacy View Security View Governance View Marketing View
      Resources Blog Read through our articles written by industry experts
      Collateral Product broch
    sentences:
      - What resources are available for learning more about GCP?
      - >-
        What are the penalties for unauthorized personal data transfer,
        including maximum fines for data fiduciaries in various scenarios?
      - What are the key provisions of South Korea's data privacy law?
pipeline_tag: sentence-similarity
model-index:
  - name: SentenceTransformer based on BAAI/bge-base-en-v1.5
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 768
          type: dim_768
        metrics:
          - type: cosine_accuracy@1
            value: 0.32989690721649484
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.5670103092783505
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.6391752577319587
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7216494845360825
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.32989690721649484
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.18900343642611683
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.12783505154639174
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.07216494845360824
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.32989690721649484
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.5670103092783505
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.6391752577319587
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7216494845360825
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.518805689291338
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.4544509900180003
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.4661116752052667
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 512
          type: dim_512
        metrics:
          - type: cosine_accuracy@1
            value: 0.3402061855670103
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.5773195876288659
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.6391752577319587
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.711340206185567
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.3402061855670103
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.1924398625429553
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.12783505154639174
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.0711340206185567
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.3402061855670103
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.5773195876288659
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.6391752577319587
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.711340206185567
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.5235302122076325
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.46329569628538714
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.4750840411397005
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 256
          type: dim_256
        metrics:
          - type: cosine_accuracy@1
            value: 0.27835051546391754
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.5154639175257731
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.5979381443298969
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7010309278350515
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.27835051546391754
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.17182130584192437
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.11958762886597937
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.07010309278350514
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.27835051546391754
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.5154639175257731
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.5979381443298969
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7010309278350515
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.4836619509866766
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.4146457208312879
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.42661551290292493
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 128
          type: dim_128
        metrics:
          - type: cosine_accuracy@1
            value: 0.31958762886597936
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.4948453608247423
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.5979381443298969
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.6804123711340206
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.31958762886597936
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.16494845360824742
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.11958762886597937
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.06804123711340206
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.31958762886597936
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.4948453608247423
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.5979381443298969
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.6804123711340206
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.48488869988900546
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.42372361315660284
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.4348164067654526
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 64
          type: dim_64
        metrics:
          - type: cosine_accuracy@1
            value: 0.25773195876288657
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.4742268041237113
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.5670103092783505
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.6494845360824743
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.25773195876288657
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.15807560137457044
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.1134020618556701
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.06494845360824741
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.25773195876288657
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.4742268041237113
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.5670103092783505
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.6494845360824743
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.4465366767058729
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.382228767795778
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.39411615598959504
            name: Cosine Map@100

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

This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5. 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: BAAI/bge-base-en-v1.5
  • Maximum Sequence Length: 512 tokens
  • Output Dimensionality: 768 tokens
  • Similarity Function: Cosine Similarity
  • Language: en
  • License: apache-2.0

Model Sources

Full Model Architecture 

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': True}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 768, '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("MugheesAwan11/bge-base-securiti-dataset-1-v13")
# Run inference
sentences = [
    "View Salesforce View Workday View GCP View Azure View Oracle View Learn more Regulations Automate compliance with global privacy regulations. US California CCPA View US California CPRA View European Union GDPR View Thailand’s PDPA View China PIPL View Canada PIPEDA View Brazil's LGPD View \\+ More View Learn more Roles Identify data risk and enable protection & control. Privacy View Security View Governance View Marketing View Resources Blog Read through our articles written by industry experts Collateral Product broch",
    'What resources are available for learning more about GCP?',
    "What are the key provisions of South Korea's data privacy law?",
]
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

Metric Value
cosine_accuracy@1 0.3299
cosine_accuracy@3 0.567
cosine_accuracy@5 0.6392
cosine_accuracy@10 0.7216
cosine_precision@1 0.3299
cosine_precision@3 0.189
cosine_precision@5 0.1278
cosine_precision@10 0.0722
cosine_recall@1 0.3299
cosine_recall@3 0.567
cosine_recall@5 0.6392
cosine_recall@10 0.7216
cosine_ndcg@10 0.5188
cosine_mrr@10 0.4545
cosine_map@100 0.4661

Information Retrieval

Metric Value
cosine_accuracy@1 0.3402
cosine_accuracy@3 0.5773
cosine_accuracy@5 0.6392
cosine_accuracy@10 0.7113
cosine_precision@1 0.3402
cosine_precision@3 0.1924
cosine_precision@5 0.1278
cosine_precision@10 0.0711
cosine_recall@1 0.3402
cosine_recall@3 0.5773
cosine_recall@5 0.6392
cosine_recall@10 0.7113
cosine_ndcg@10 0.5235
cosine_mrr@10 0.4633
cosine_map@100 0.4751

Information Retrieval

Metric Value
cosine_accuracy@1 0.2784
cosine_accuracy@3 0.5155
cosine_accuracy@5 0.5979
cosine_accuracy@10 0.701
cosine_precision@1 0.2784
cosine_precision@3 0.1718
cosine_precision@5 0.1196
cosine_precision@10 0.0701
cosine_recall@1 0.2784
cosine_recall@3 0.5155
cosine_recall@5 0.5979
cosine_recall@10 0.701
cosine_ndcg@10 0.4837
cosine_mrr@10 0.4146
cosine_map@100 0.4266

Information Retrieval

Metric Value
cosine_accuracy@1 0.3196
cosine_accuracy@3 0.4948
cosine_accuracy@5 0.5979
cosine_accuracy@10 0.6804
cosine_precision@1 0.3196
cosine_precision@3 0.1649
cosine_precision@5 0.1196
cosine_precision@10 0.068
cosine_recall@1 0.3196
cosine_recall@3 0.4948
cosine_recall@5 0.5979
cosine_recall@10 0.6804
cosine_ndcg@10 0.4849
cosine_mrr@10 0.4237
cosine_map@100 0.4348

Information Retrieval

Metric Value
cosine_accuracy@1 0.2577
cosine_accuracy@3 0.4742
cosine_accuracy@5 0.567
cosine_accuracy@10 0.6495
cosine_precision@1 0.2577
cosine_precision@3 0.1581
cosine_precision@5 0.1134
cosine_precision@10 0.0649
cosine_recall@1 0.2577
cosine_recall@3 0.4742
cosine_recall@5 0.567
cosine_recall@10 0.6495
cosine_ndcg@10 0.4465
cosine_mrr@10 0.3822
cosine_map@100 0.3941

Training Details

Training Dataset

Unnamed Dataset

  • Size: 872 training samples
  • Columns: positive and anchor
  • Approximate statistics based on the first 1000 samples:
    positive anchor
    type string string
    details
    • min: 89 tokens
    • mean: 229.38 tokens
    • max: 414 tokens
    • min: 9 tokens
    • mean: 21.92 tokens
    • max: 102 tokens
  • Samples:
    positive anchor
    controller should inform the data subject in every situation where his or her personal data is processed. The LPPD provides a general requirement to provide information on the collection methods but does not explicitly refer to automated decision-making or profiling. vs Articles: 5 14, Recitals: 58 63 This right requires the controller to provide the following information to the data subject when requested. This should be given in a concise, transparent, intelligible, and easily accessible form, using plain language: The identity and contact details of the controller, controller’s representative, and DPO, where applicable The purpose and the legal basis of the processing The categories of personal data concerned The recipients of the personal data The appropriate or suitable safeguards and the means to obtain a copy of them or where they have been made available The controller must provide information necessary to ensure fair and transparent processing whether or not the personal What information must the controller provide regarding their identity and contact details?
    and deletions, and manage all vendor contracts and compliance documents. ## Key Rights Under Ghana’s Data Protection Act 2012 Right to be Informed : Data subjects have the right to be informed of the processing of their personal data and the purposes for which the data is processed. Right to Access: Data subjects have the right to obtain confirmation whether or not the controller holds personal data about them, access their personal data, and obtain descriptions of data recipients. Right to Rectification : Under the right to rectification, data subjects can request the correction of their data. Right to Erasure: Data subjects have the right to request the erasure and destruction of the data that is no longer needed by the organization. Right to Object: The data subject has the right to prevent the data controller from processing personal data if such processing causes or is likely to cause unwarranted damage or distress to the data What are the key rights provided to data subjects under Ghana's Data Protection Act 2012?
    aim to protect personal data, they have differences in scope, requirements, and applicability. PDPA applies to Thailand, while GDPR applies to the European Union. The effect of PDPA in Thailand is to regulate how personal data is processed, collected, used, and protected by individuals and organizations in the country. Thailand's PDPA includes provisions related to personal data breach notifications, requiring data controllers to notify the Personal Data Protection Committee (PDPC) of a personal data breach as soon as possible, preferably within 72 hours of becoming aware of it. The principles of PDPA in Thailand include obtaining consent, especially for minors, ensuring data security, issuing timely data breach notifications, designating a data protection officer, conducting data protection impact assessments, maintaining a record of processing activities, and ensuring adequate standards when transferring data across borders. ## Join Our Newsletter Get all the latest information, law updates and more delivered to your inbox ### Share What is the role of obtaining consent in Thailand's PDPA?
  • Loss: MatryoshkaLoss with these parameters:
    {
    "loss": "MultipleNegativesRankingLoss",
    "matryoshka_dims": [
        768,
        512,
        256,
        128,
        64
    ],
    "matryoshka_weights": [
        1,
        1,
        1,
        1,
        1
    ],
    "n_dims_per_step": -1
}

Training Hyperparameters

Non-Default Hyperparameters

  • eval_strategy: epoch
  • per_device_train_batch_size: 32
  • per_device_eval_batch_size: 16
  • learning_rate: 2e-05
  • num_train_epochs: 2
  • lr_scheduler_type: cosine
  • warmup_ratio: 0.1
  • bf16: True
  • tf32: True
  • load_best_model_at_end: True
  • optim: adamw_torch_fused
  • batch_sampler: no_duplicates

All Hyperparameters

Click to expand
  • overwrite_output_dir: False
  • do_predict: False
  • eval_strategy: epoch
  • prediction_loss_only: True
  • per_device_train_batch_size: 32
  • per_device_eval_batch_size: 16
  • per_gpu_train_batch_size: None
  • per_gpu_eval_batch_size: None
  • gradient_accumulation_steps: 1
  • eval_accumulation_steps: None
  • learning_rate: 2e-05
  • weight_decay: 0.0
  • adam_beta1: 0.9
  • adam_beta2: 0.999
  • adam_epsilon: 1e-08
  • max_grad_norm: 1.0
  • num_train_epochs: 2
  • 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: True
  • fp16: False
  • fp16_opt_level: O1
  • half_precision_backend: auto
  • bf16_full_eval: False
  • fp16_full_eval: False
  • tf32: True
  • 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: True
  • 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_fused
  • 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: False
  • hub_always_push: False
  • gradient_checkpointing: False
  • gradient_checkpointing_kwargs: None
  • include_inputs_for_metrics: False
  • 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
  • batch_sampler: no_duplicates
  • multi_dataset_batch_sampler: proportional

Training Logs

Epoch Step Training Loss dim_128_cosine_map@100 dim_256_cosine_map@100 dim_512_cosine_map@100 dim_64_cosine_map@100 dim_768_cosine_map@100
0.3571 10 6.4098 - - - - -
0.7143 20 4.9339 - - - - -
1.0 28 - 0.4266 0.4263 0.4703 0.3934 0.4650
1.0714 30 3.7606 - - - - -
1.4286 40 2.5546 - - - - -
1.7857 50 3.1845 - - - - -
2.0 56 - 0.4348 0.4266 0.4751 0.3941 0.4661
  • The bold row denotes the saved checkpoint.

Framework Versions

  • Python: 3.10.14
  • Sentence Transformers: 3.0.1
  • Transformers: 4.41.2
  • PyTorch: 2.1.2+cu121
  • Accelerate: 0.31.0
  • Datasets: 2.19.1
  • Tokenizers: 0.19.1

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",
}

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}
}