axondendriteplus's picture
Update README.md
8228570 verified
metadata
language:
  - en
license: apache-2.0
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:1456
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
base_model: Snowflake/snowflake-arctic-embed-m-v2.0
widget:
  - source_sentence: >-
      relation to a portfolio manager being a body corporate, shall be construed
      with reference to : (i) the definition of control in terms of Regulation
      2(1)(e) of SEBI (Substantial Acquisition of Shares and Takeovers)
      Regulations, 2011 as amended from time to time, if its shares are listed
      on any recognized stock exchange; (ii) in any other case, change in the
      controlling interest in the body corporate; Explanation. For the purpose
      of sub-clause (ii), the expression controlling interest means, (A) an
      interest, whether direct or indirect, to the extent of at least fifty-one
      percent of voting rights in the body corporate; (B) right to appoint
      majority of the directors or to control the management directly or
      indirectly. Page 3 of 78 reference to the definition of control in terms
      of regulations framed under clause (h) of sub-section (2) of section 11 of
      the Act; (B) if its shares are not listed on any recognised stock
      exchange, shall be construed with reference to the definition of control
      as provided in sub-section (27) of Section 2 of the Companies Act, 2013
      (18 of 2013);] (f) chartered accountant" means a chartered accountant as
      defined in clause (b) of sub- section (1) of section 2 of the Chartered
      Accountants Act, 1949 (38 of 1949) and who has obtained a certificate of
      practice under sub-section (1) of section 6 of that Act; 5[(fa)
      Co-investment Portfolio Manager means a Portfolio Manager who is a Manager
      of a Category I or Category II Alternative Investment Fund(s); and: (i)
      provides services only to the investors of such Category I or Category II
      Alternative Investment Fund(s); and (ii) makes investment only in unlisted
      securities of investee companies where such Category I or Category II
      Alternative Investment Fund(s) make investments: Provided that the
      Co-investment Portfolio Manager may provide services to investors from any
      other Category I or Category II Alternative Investment Fund(s) which are
      managed by them and are also sponsored by the same Sponsor(s);] (g)
      discretionary portfolio manager means a portfolio manager who under a
      contract relating to portfolio management, exercises or may exercise, any
      degree of discretion as to the investment of funds or management of the
      portfolio of securities of the client, as the case may be; (h) eligible
      fund manager shall have the same meaning as assigned to it in sub-section
      (4) of Section 9A of the Income-tax Act, 1961; (i) eligible investment
      fund shall have the same meaning as assigned to it in sub-section (3) of
      Section 9A of the Income-tax Act, 1961; 5 Inserted by the Securities and
      Exchange Board of India (Portfolio Managers) (Fourth Amendment)
      Regulations, 2021 w. Page 4 of 78 (j) "form" means a form specified in
      Schedule I; (k) goods means the goods notified by the Central Government
      under clause (bc) of section 2 of the Securities Contracts (Regulation)
      Act, 1956 and forming the underlying of any commodity derivative; (l)
      "inspecting authority" means one or more persons appointed by the Board to
      exercise powers conferred under Chapter V; 6[(la) large value accredited
      investor means an accredited investor who has entered into an agreement
      with the portfolio manager for a minimum investment amount of ten crore
      rupees;] 7[(lb) investee company shall have the same meaning as assigned
      to it in clause (o) of sub- regulation (1) of regulation 2 of the
      Securities and Exchange Board of India (Alternative Investment Funds)
      Regulations, 2012; (lc) Manager shall
    sentences:
      - >-
        What are the particulars of units of the scheme and/or shares and
        debentures of the company issued for consideration other than cash?
      - >-
        What is the maximum period allowed for a listed company to increase its
        public shareholding to twenty-five percent after it falls below that
        threshold as a result of an approved resolution plan under the
        Insolvency and Bankruptcy Code, 2016?
      - >-
        What is the definition of "controlling interest" in the context of a
        body corporate under the SEBI regulations?
  - source_sentence: >-
      The issuer making a private placement of debt securities and
      non-convertible redeemable preference shares and seeking listing thereof
      on a recognised stock exchange shall make the following disclosures in the
      placement memorandum: (a) disclosures specified in 28[Schedule I] of these
      regulations; (b) disclosures specified in the Companies Act, 2013 (18 of
      2013), as applicable; (c) additional disclosures as may be specified by
      the Board. (2) The disclosures as provided in sub-regulation (1) shall be
      made on the websites of stock exchange(s) where such securities are
      proposed to be listed and shall be available for download in PDF or any
      other format as may be specified by the Board. (3) The issuer shall ensure
      that the audited financial statements contained in the placement
      memorandum 29[] shall not be more than six months old from the date of
      filing placement memorandum or the issue opening date, as applicable: 27
      Substituted by the by the Securities and Exchange Board of India (Issue
      and Listing of Non-Convertible Securities) (Amendment) Regulations, 2024
      w. Prior to substitution, it read as The debenture trustee shall submit a
      due diligence certificate to the stock exchange: (a) in case of secured
      debt securities, in the format as specified in Schedule IV of these
      regulations; and (b) in case of unsecured debt securities, in the format
      as specified in Schedule IVA of these regulations. Prior to this, it was
      substituted by the Securities and Exchange Board of India (Issue and
      Listing of Non-Convertible Securities) (Amendment) Regulations, 2022, w.
      Prior to substitution, sub-regulation 3 read as: Debenture trustee shall
      submit a due diligence certificate to the stock exchange in the format as
      specified in Schedule IV of these regulations. 28 Substituted by the
      Securities and Exchange Board of India (Issue and Listing of
      Non-Convertible Securities) (Second Amendment) Regulations, 2023, w. Prior
      to substitution, the words were Schedule II. 29 Omitted by the Securities
      and Exchange Board of India (Issue and Listing of Non-Convertible
      Securities) (Second Amendment) Regulations, 2023, w. Prior to omission,
      the words were and tranche placement memorandum. Page 26 of 79 Provided
      that in case of: (a) listed issuers (whose non-convertible securities or
      specified securities are listed on recognised stock exchange(s)), who are
      in compliance with the listing regulations; (b) the issuers of
      non-convertible securities, who are subsidiaries of entities who have
      listed their specified securities, and are in compliance with the listing
      regulations, instead of audited financial statements for the stub period,
      they may disclose unaudited financial information for such period in the
      format as prescribed in the listing regulations with limited review
      report, as filed with the stock exchange(s), subject to necessary
      disclosures in this regard in the placement memorandum including risk
      factors. Allotment of securities 46. The issuer shall ensure allotment of
      debt securities and non-convertible redeemable preference shares issued on
      a private placement basis and credit to the dematerialised account of the
      investors, is made within such time as may be specified by the Board. PART
      B ADDITIONAL PROVISIONS FOR LISTING OF DEBT SECURITIES ISSUED ON PRIVATE
      PLACEMENT BASIS Filing of shelf placement memorandum 47. 30[] Creation of
      security 48. (1) While creating a charge or security, the issuer shall
      have the option to create charge or security over the properties or assets
      (movable, immovable, tangible, intangible), shares or any interest
      thereon, of the issuer or its subsidiaries or its holding companies or its
      associate companies. 30 Omitted by the Securities and Exchange Board of
      India
    sentences:
      - >-
        What is the penalty for a reporting requirement violation under the PIT
        Regulations if there is a delay of more than three months?
      - >-
        What are the names of the members of the issuer's audit committee,
        nomination and remuneration committee, and stakeholders relationship
        committee?
      - >-
        What disclosures must an issuer include in the placement memorandum when
        making a private placement of debt securities and non-convertible
        redeemable preference shares?
  - source_sentence: >-
      Accredited refineries means refineries empanelled by the Stock Exchanges;
      Page 4 of 26 (c) Assayer means a person engaged in the process of
      assessing the purity or quality of gold; (d) Beneficial Owner means a
      person whose name is recorded as such with a depository/depository
      participant; (e) Board means the Securities and Exchange Board of India
      established under section 3 of the Act; 1[(f) Change in control in case of
      a body corporate (A) if its shares are listed on any recognised stock
      exchange, shall be construed with reference to the definition of control
      in terms of regulations framed under clause (h) of sub-section (2) of
      section 11 of the Act; (B) if its shares are not listed on any recognised
      stock exchange, shall be construed with reference to the definition of
      control as provided in sub-section (27) of Section 2 of the Companies Act,
      2013 (18 of 2013);] (g) Depositor means a person who owns the gold
      deposited with the vault for creation of Electronic Gold Receipt and its
      trading on recognized stock exchange; (h) Electronic Gold Receipt shall
      have the meaning assigned to it under the Securities Contracts
      (Regulation) Act, 1956; (i) Gold standard means the purity and standard of
      gold as specified by the recognized stock exchanges; (j) Nominated
      agencies means agencies nominated by the Directorate General of Foreign
      Trade for import of Gold under the Foreign Trade (Development and
      Regulation) Act, 1992; 1 Substituted by the Securities and Exchange Board
      of India (Change in Control in Intermediaries) (Amendment) Regulations,
      2023 w. Prior to the substitution, clause (f) read as under: Change in
      control, in relation to a Vault Manager being a body corporate, shall be
      construed with reference to: (i) the definition of control in terms of
      regulation 2(1)(e) of SEBI (Substantial Acquisition of Shares and
      Takeovers) Regulations, 2011 as amended from time to time if its shares
      are listed on any recognized stock exchange; (ii) in any other case,
      change in the controlling interest in the body corporate; Explanation For
      the purpose of sub-clause (ii), the expression controlling interest means-
      (A) an interest, whether direct or indirect, to the extent of at least
      fifty-one percent of voting rights in the body corporate; or (B) right to
      appoint majority of the directors or to control the management directly or
      indirectly; Page 5 of 26 (k) Recognized vault means the premises
      encompassing strong room(s) set up and managed by the Vault Manager and
      which conforms with all the requirements specified by the Board for the
      purpose of providing vaulting services; (l) Vault Manager means any person
      who carries on or intends to carry on the business of providing vaulting
      services; (m) Vaulting service in relation to gold means the storage and
      safekeeping of gold deposited with the Vault Manager, by the depositor,
      for the purpose of trading in Electronic Gold Receipt and providing
      services incidental thereto, and includes (i) utilizing the services of
      assayers empanelled with the Stock Exchanges for testing as per the gold
      standard, wherever required; (ii) coordination with depositories for
      creation, transfer and extinguishment of Electronic Gold Receipt; and
      (iii) providing deposit, storage and withdrawal services to the beneficial
      owners. (2) The words and expressions used and not defined in these
      regulations, but defined in the Act, the Securities Contracts (Regulation)
      Act, 1956, (42 of 1956), the Companies Act, 2013 (18 of 2013), the
      Depositories Act, 1996, or any rules
    sentences:
      - >-
        What happens if the eighth day, as reckoned under the Negotiable
        Instruments Act, is itself a public holiday?
      - What is the role of an Assayer according to the context provided?
      - >-
        What disclosures are required regarding significant income sources that
        constitute more than 10% of total income?
  - source_sentence: >-
      Regulations, 2023 w. 93 Inserted by the Securities and Exchange Board of
      India (Buy-Back of Securities) (Second Amendment) Regulations, 2024 w.
      Particulars Content Public Announcement i) The Public announcement shall
      be dated and signed on behalf of the Board of Directors of the company by
      its manager or secretary, if any, and by not less than two directors of
      the company one of whom shall be a managing director where there is one.
      ii) A full and complete disclosure of all material facts including the
      disclosures mentioned in Schedule I. iii) In addition to the disclosures
      in Schedule A, the following disclosures shall be made: i) Date of
      shareholders approval for buy-back, if applicable; ii) Minimum and maximum
      number of securities that the company proposes to buy-back, sources of
      funds from which the buy-back would be made and the cost of financing the
      buy-back; iii) Proposed time table from opening of offer till the
      extinguishment of the certificates; iv) Process and methodology to be
      adopted for the buy- back; v) Brief information about the company; i) The
      Public announcement shall be dated and signed on behalf of the Board of
      Directors of the company by its manager or secretary, if any, and by not
      less than two directors of the company one of whom shall be a managing
      director where there is one. ii) A full and complete disclosure of all
      material facts including the disclosures mentioned in Schedule I. Page 46
      of 51 SCHEDULE - IV [Regulation 16(iv)(b)] Public Announcement for Open
      Market Buy-Back through Stock Exchange Particulars Content Public
      Announcement i) The Public announcement shall be dated and signed on
      behalf of the Board of Directors of the company by its manager or
      secretary, if any, and by not less than two directors of the company one
      of whom shall be a managing director where there is one. ii) A full and
      complete disclosure of all material facts including the disclosures
      mentioned in Schedule I. iii) In addition to the disclosures in Schedule
      A, the following disclosures shall be made: i) Date of shareholders
      approval for buy-back, if applicable; ii) Minimum and maximum number of
      securities that the company proposes to buy-back, sources of funds from
      which the buy-back would be made and the cost of financing the buy-back;
      iii) Proposed time table from opening of offer till the extinguishment of
      the certificates; iv) Process and methodology to be adopted for the buy-
      back; v) Brief information about the company; Particulars Content vi)
      Audited Financial information for the last 3 years and the lead manager
      shall ensure that the particulars (audited statement and un-audited
      statement) contained therein shall not be more than more than 6 months old
      from the date of the public announcement together with financial ratios as
      may be specified by the Board; Explanation: Ensure that the un-audited
      financial results, if any disclosed, should be certified / limited review
      by statutory auditors. vii) Details of escrow account opened and the
      amount deposited therein; viii) Listing details and stock market data: a)
      high, low and average market prices of the securities of the company
      proposed to be bought back, during the preceding three years; b) monthly
      high and low prices for the six months preceding the date of the public
      announcement; c) the number of securities traded on the days when the high
      and low prices were recorded on the relevant stock exchanges during the
      period stated at (a) and (b) above; d) the stock market data referred to
      above shall be shown separately for periods marked by a change in capital
      structure, with such period commencing from the date the concerned stock
      exchange recognises the change in the capital structure. when the
      securities have become ex-rights or ex-bonus) ; e)
    sentences:
      - >-
        What is the time frame within which a return of allotment of securities
        must be filed with the Registrar according to the Companies
        (Registration Offices and Fees) Rules, 2014?
      - >-
        What is the deadline for a listed entity to put in place systems and
        processes for compliance with clause (f) of sub-regulation (2) of
        regulation 34 after it is required to comply for the first time?
      - >-
        What are the specific disclosures required in a public announcement for
        a buy-back of securities according to the Securities and Exchange Board
        of India regulations?
  - source_sentence: >-
      1996. In section 19 of the Depositories Act, 1996 (hereafter in this
      chapter referred to as the principal Act in this chapter), the following
      Explanation shall be inserted, namely:- Explanation. For the removal of
      doubts, it is hereby declared that power to issue directions under this
      section shall include and always be deemed to have been included the power
      to direct any person, who made profit or averted loss by indulging in any
      transaction or activity in contravention of the provisions of this Act or
      regulations made thereunder, to disgorge an amount equivalent to the
      wrongful gain made or loss averted by such contravention.
    sentences:
      - >-
        What are the qualifications required for a judge to be appointed to a
        Special Court under this Act?
      - >-
        What is the purpose of the Explanation inserted in section 19 of the
        Depositories Act, 1996?
      - >-
        What is the minimum percentage of total issued shares that the acquirer
        must reach in order for the delisting offer to be considered successful?
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: BGE base Financial Matryoshka
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 768
          type: dim_768
        metrics:
          - type: cosine_accuracy@1
            value: 0.4691358024691358
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7469135802469136
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.845679012345679
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.9135802469135802
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.4691358024691358
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.24897119341563784
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.1691358024691358
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.09135802469135802
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.4691358024691358
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7469135802469136
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.845679012345679
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.9135802469135802
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6931888396302245
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6220483049186752
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6258770441533618
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 512
          type: dim_512
        metrics:
          - type: cosine_accuracy@1
            value: 0.4506172839506173
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7530864197530864
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.8271604938271605
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.9074074074074074
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.4506172839506173
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.2510288065843621
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.16543209876543208
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.09074074074074073
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.4506172839506173
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7530864197530864
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.8271604938271605
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.9074074074074074
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6862124164896819
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6141950813247109
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.618139737272647
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 256
          type: dim_256
        metrics:
          - type: cosine_accuracy@1
            value: 0.4382716049382716
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7345679012345679
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.8271604938271605
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8950617283950617
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.4382716049382716
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.24485596707818924
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.16543209876543208
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.08950617283950617
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.4382716049382716
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7345679012345679
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.8271604938271605
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8950617283950617
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6725732937028854
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6002816970409561
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6055931590500198
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 128
          type: dim_128
        metrics:
          - type: cosine_accuracy@1
            value: 0.41358024691358025
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.6851851851851852
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.7777777777777778
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8703703703703703
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.41358024691358025
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.22839506172839505
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.15555555555555553
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.08703703703703702
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.41358024691358025
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.6851851851851852
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.7777777777777778
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8703703703703703
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6396728651848874
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.5658534195571232
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.5718872882660476
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 64
          type: dim_64
        metrics:
          - type: cosine_accuracy@1
            value: 0.345679012345679
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.5802469135802469
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.6851851851851852
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7901234567901234
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.345679012345679
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.19341563786008228
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.13703703703703704
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.07901234567901233
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.345679012345679
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.5802469135802469
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.6851851851851852
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7901234567901234
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.5603350026673091
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.48750489907897315
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.49429085785187665
            name: Cosine Map@100

Legal-Embed-snowflake-arctic-embed-m-v2.0

This is a sentence-transformers model finetuned from Snowflake/snowflake-arctic-embed-m-v2.0 on the legal-rag-embedding-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
  • Base model: Snowflake/snowflake-arctic-embed-m-v2.0
  • Maximum Sequence Length: 8192 tokens
  • Output Dimensionality: 768 dimensions
  • Similarity Function: Cosine Similarity
  • Training Dataset:
    • json
  • Language: en
  • License: apache-2.0

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("axondendriteplus/Legal-Embed-snowflake-arctic-embed-m-v2.0")
# Run inference
sentences = [
    '1996. In section 19 of the Depositories Act, 1996 (hereafter in this chapter referred to as the principal Act in this chapter), the following Explanation shall be inserted, namely:- Explanation. For the removal of doubts, it is hereby declared that power to issue directions under this section shall include and always be deemed to have been included the power to direct any person, who made profit or averted loss by indulging in any transaction or activity in contravention of the provisions of this Act or regulations made thereunder, to disgorge an amount equivalent to the wrongful gain made or loss averted by such contravention.',
    'What is the purpose of the Explanation inserted in section 19 of the Depositories Act, 1996?',
    'What are the qualifications required for a judge to be appointed to a Special Court under this Act?',
]
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 0.4691
cosine_accuracy@3 0.7469
cosine_accuracy@5 0.8457
cosine_accuracy@10 0.9136
cosine_precision@1 0.4691
cosine_precision@3 0.249
cosine_precision@5 0.1691
cosine_precision@10 0.0914
cosine_recall@1 0.4691
cosine_recall@3 0.7469
cosine_recall@5 0.8457
cosine_recall@10 0.9136
cosine_ndcg@10 0.6932
cosine_mrr@10 0.622
cosine_map@100 0.6259

Information Retrieval

Metric Value
cosine_accuracy@1 0.4506
cosine_accuracy@3 0.7531
cosine_accuracy@5 0.8272
cosine_accuracy@10 0.9074
cosine_precision@1 0.4506
cosine_precision@3 0.251
cosine_precision@5 0.1654
cosine_precision@10 0.0907
cosine_recall@1 0.4506
cosine_recall@3 0.7531
cosine_recall@5 0.8272
cosine_recall@10 0.9074
cosine_ndcg@10 0.6862
cosine_mrr@10 0.6142
cosine_map@100 0.6181

Information Retrieval

Metric Value
cosine_accuracy@1 0.4383
cosine_accuracy@3 0.7346
cosine_accuracy@5 0.8272
cosine_accuracy@10 0.8951
cosine_precision@1 0.4383
cosine_precision@3 0.2449
cosine_precision@5 0.1654
cosine_precision@10 0.0895
cosine_recall@1 0.4383
cosine_recall@3 0.7346
cosine_recall@5 0.8272
cosine_recall@10 0.8951
cosine_ndcg@10 0.6726
cosine_mrr@10 0.6003
cosine_map@100 0.6056

Information Retrieval

Metric Value
cosine_accuracy@1 0.4136
cosine_accuracy@3 0.6852
cosine_accuracy@5 0.7778
cosine_accuracy@10 0.8704
cosine_precision@1 0.4136
cosine_precision@3 0.2284
cosine_precision@5 0.1556
cosine_precision@10 0.087
cosine_recall@1 0.4136
cosine_recall@3 0.6852
cosine_recall@5 0.7778
cosine_recall@10 0.8704
cosine_ndcg@10 0.6397
cosine_mrr@10 0.5659
cosine_map@100 0.5719

Information Retrieval

Metric Value
cosine_accuracy@1 0.3457
cosine_accuracy@3 0.5802
cosine_accuracy@5 0.6852
cosine_accuracy@10 0.7901
cosine_precision@1 0.3457
cosine_precision@3 0.1934
cosine_precision@5 0.137
cosine_precision@10 0.079
cosine_recall@1 0.3457
cosine_recall@3 0.5802
cosine_recall@5 0.6852
cosine_recall@10 0.7901
cosine_ndcg@10 0.5603
cosine_mrr@10 0.4875
cosine_map@100 0.4943

Training Details

Training Dataset

json

  • Dataset: json
  • Size: 1,456 training samples
  • Columns: positive and anchor
  • Approximate statistics based on the first 1000 samples:
    positive anchor
    type string string
    details
    • min: 110 tokens
    • mean: 795.06 tokens
    • max: 1042 tokens
    • min: 13 tokens
    • mean: 30.5 tokens
    • max: 215 tokens
  • Samples:
    positive anchor
    its continued obligations towards the holders of debt securities. We have satisfied ourselves about the ability of the issuer to service the debt securities. PLACE DATE: DEBENTURE TRUSTEE TO THE ISSUE WITH HIS SEAL Page - 65 - of 68 FORMAT OF DUE DILIGENCE CERTIFICATE TO BE GIVEN BY THE DEBENTURE TRUSTEE BEFORE OPENING OF THE ISSUE To, SECURITIES AND EXCHANGE BOARD OF INDIA Dear Sir / Madam, SUB. : ISSUE OF BY (Issuer) We, the Debenture Trustee (s) to the above mentioned forthcoming issue state as follows: (1) We have examined documents pertaining to the said issue and other relevant documents. (2) On the basis of such examination and discussions with the issuer, its Mayor/Deputy Mayor /Directors and other officers, other agencies and independent verification of the various relevant documents,- (a) WE CONFIRM that the issuer has made adequate provisions regarding escrow payment mechanism for repayment of debt obligations, and (b) We have satisfied ourselves about the ability of the iss... What specific provisions has the issuer made regarding the repayment of debt obligations?
    sums realised by way of penalties to Consolidated Fund of India 23L. Appeal to Securities Appellate Tribunal 23M. Offences 23N. Composition of certain offences 23-O. Power to grant immunity 24. Contravention by companies 25. Certain offences to be cognizable 26. Cognizance of offences by courts 26A. Establishment of Special Courts 26B. Offences triable by Special Courts 26C. Appeal and revision 26D. Application of Code to proceedings before Special Court 26E. Transitional Provisions MISCELLANEOUS 27. Title to dividends 27A. Right to receive income from collective investment scheme 27B. Right to receive income from mutual fund 28. Act not to apply in certain cases 29. Protection of action taken in good faith 29A. Power to delegate 29B. Powers of Board not to apply to International Financial Services Centre 30. Power to make rules 30A. Special Provisions related to commodity derivatives 30B. Special provisions related to pooled investment vehicle 31. Power of Securities and Exchange Boar... What powers does the Securities and Exchange Board of India have to make regulations according to the Securities Contracts (Regulation) Act, 1956?
    the depository or the securities market as a result of the default; and (c) the repetitive nature of the default. ] CHAPTER X PROCEDURE FOR ACTION IN CASE OF DEFAULT Liability for action in case of default 92. Without prejudice to the power of the Board to take action, under the provisions of the Act and the Depositories Act, if a depository or a participant:- (a) contravenes any of the provisions of the Act, the Depositories Act, the bye-laws, agreements and these regulations; (b) fails to furnish any information relating to its activity as a depository or participant as required under these regulations; (c) does not furnish the information called for by the Board under clause (a) of sub-section (1) of section 18 of the Depositories Act or furnishes information which is false or misleading in any material particular; (d) does not co-operate in any inspection or investigation or enquiry conducted by the Board; (e) fails to comply with any direction of the Board issued under section 18 ... What actions can the Board take against a depository or participant that fails to comply with the provisions of the Act or the Depositories Act?
  • 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
  • gradient_accumulation_steps: 16
  • learning_rate: 2e-05
  • num_train_epochs: 4
  • 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: 16
  • eval_accumulation_steps: None
  • torch_empty_cache_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: 4
  • 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: 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: no_duplicates
  • multi_dataset_batch_sampler: proportional

Training Logs

Epoch Step Training Loss dim_768_cosine_ndcg@10 dim_512_cosine_ndcg@10 dim_256_cosine_ndcg@10 dim_128_cosine_ndcg@10 dim_64_cosine_ndcg@10
1.0 3 - 0.6688 0.6577 0.6456 0.6018 0.5202
2.0 6 - 0.6652 0.6624 0.6634 0.6147 0.5320
3.0 9 - 0.6868 0.6831 0.6678 0.6330 0.5537
3.3478 10 39.9839 - - - - -
4.0 12 - 0.6932 0.6862 0.6726 0.6397 0.5603
  • The bold row denotes the saved checkpoint.

Framework Versions

  • Python: 3.11.11
  • Sentence Transformers: 4.1.0
  • Transformers: 4.52.3
  • PyTorch: 2.7.0+cu126
  • Accelerate: 1.7.0
  • 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",
}

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}
}