Tejasw1/gte-base-legal-matryoshka-v1

SentenceTransformer based on Alibaba-NLP/gte-base-en-v1.5

This is a sentence-transformers model finetuned from Alibaba-NLP/gte-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: Alibaba-NLP/gte-base-en-v1.5
• Maximum Sequence Length: 8192 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity
• Language: en
• License: apache-2.0

Model Sources

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

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 8192, 'do_lower_case': False}) with Transformer model: NewModel 
  (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})
)

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("Tejasw1/gte-base-legal-matryoshka-v1")
# Run inference
sentences = [
    '(1) Save as otherwise provided in the Insolvency and Bankruptcy Code, 2016 (31 of 2016), where a person, liable to pay tax, interest or penalty under this Act, dies, then (a) if a business carried on by the person is continued after his death by his legal representative or any other person, such legal representative or other person, shall be liable to pay tax, interest or penalty due from such person under this Act; and (b) if the business carried on by the person is discontinued, whether before or after his death, his legal representative shall be liable to pay, out of the estate of the deceased, to the extent to which the estate is capable of meeting the charge, the tax, interest or penalty due from such person under this Act, whether such tax, interest or penalty has been determined before his death but has remained unpaid or is determined after his death. (2) Save as otherwise provided in the Insolvency and Bankruptcy Code, 2016 (31 of 2016), where a taxable person, liable to pay tax, interest or penalty under this Act, is a Hindu Undivided Family or an association of persons and the property of the Hindu Undivided Family or the association of persons is partitioned amongst the various members or groups of members, then, each member or group of members shall, jointly and severally, be liable to pay the tax, interest or penalty due from the taxable person under this Act up to the time of the partition whether such tax, penalty or interest has been determined before partition but has remained unpaid or is determined after the partition. (3) Save as otherwise provided in the Insolvency and Bankruptcy Code, 2016 (31 of 2016), where a taxable person, liable to pay tax, interest or penalty under this Act, is a firm, and the firm is dissolved, then, every person who was a partner shall, jointly and severally, be liable to pay the tax, interest or penalty due from the firm under this Act up to the time of dissolution whether such tax, interest or penalty has been determined before the dissolution, but has remained unpaid or is determined after dissolution. (4) Save as otherwise provided in the Insolvency and Bankruptcy Code, 2016 (31 of 2016), where a taxable person liable to pay tax, interest or penalty under this Act,-- (a) is the guardian of a ward on whose behalf the business is carried on by the guardian; or (b) is a trustee who carries on the business under a trust for a beneficiary, then, if the guardianship or trust is terminated, the ward or the beneficiary shall be liable to pay the tax, interest or penalty due from the taxable person upto the time of the termination of the guardianship or trust, whether such tax, interest or penalty has been determined before the termination of guardianship or trust but has remained unpaid or is determined thereafter.',
    'Under the Central Goods and Services Tax Act, 2017, what are the tax liabilities of a legal representative if a business continues after the death of a person liable to pay tax, interest, or penalty?',
    'According to The Insolvency and Bankruptcy Code, 2016, Section 160(2), can a bankruptcy trustee give a notice of disclaimer for onerous property even if they have taken possession of it or tried to sell it?',
]
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: dim_768
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.0193
cosine_accuracy@3	0.1213
cosine_accuracy@5	0.3456
cosine_accuracy@10	0.614
cosine_precision@1	0.0193
cosine_precision@3	0.0404
cosine_precision@5	0.0691
cosine_precision@10	0.0614
cosine_recall@1	0.0193
cosine_recall@3	0.1213
cosine_recall@5	0.3456
cosine_recall@10	0.614
cosine_ndcg@10	0.2539
cosine_mrr@10	0.1458
cosine_map@100	0.1624

Information Retrieval

• Dataset: dim_512
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.0129
cosine_accuracy@3	0.125
cosine_accuracy@5	0.3511
cosine_accuracy@10	0.6176
cosine_precision@1	0.0129
cosine_precision@3	0.0417
cosine_precision@5	0.0702
cosine_precision@10	0.0618
cosine_recall@1	0.0129
cosine_recall@3	0.125
cosine_recall@5	0.3511
cosine_recall@10	0.6176
cosine_ndcg@10	0.2532
cosine_mrr@10	0.1437
cosine_map@100	0.1594

Training Details

Training Dataset

Unnamed Dataset

• Size: 9,791 training samples
• Columns: positive and anchor
• Approximate statistics based on the first 1000 samples:

  	positive	anchor
  type	string	string
  details	min: 16 tokens mean: 769.8 tokens max: 7913 tokens	min: 13 tokens mean: 32.89 tokens max: 77 tokens

• Samples:

  positive	anchor
  A party to any proceeding or appeal before the Tribunal or the Appellate Tribunal, as the case may be, may either appear in person or authorise one or more chartered accountants or company secretaries or cost accountants or legal practitioners or any other person to present his case before the Tribunal or the Appellate Tribunal, as the case may be	According to the Companies Act, 2013, Section 432, are company secretaries allowed to represent a party in proceedings before the Tribunal?
  Show Related Subordinates (1) There shall be filed with the Registrar within whose jurisdiction the registered office of a company is proposed to be situated, the following documents and information for registration, namely:-- (a) the memorandum and articles of the company duly signed by all the subscribers to the memorandum in such manner as may be prescribed; (b) a declaration in the prescribed form by an advocate, a chartered accountant, cost accountant or company secretary in practice, who is engaged in the formation of the company, and by a person named in the articles as a director, manager or secretary of the company, that all the requirements of this Act and the rules made thereunder in respect of registration and matters precedent or incidental thereto have been complied with; (c) 1[a declaration] from each of the subscribers to the memorandum and from persons named as the first directors, if any, in the articles that he is not convicted of any offence in connection with the promotion, formation or management of any company, or that he has not been found guilty of any fraud or misfeasance or of any breach of duty to any company under this Act or any previous company law during the preceding five years and that all the documents filed with the Registrar for registration of the company contain information that is correct and complete and true to the best of his knowledge and belief; (d) the address for correspondence till its registered office is established; (e) the particulars of name, including surname or family name, residential address, nationality and such other particulars of every subscriber to the memorandum along with proof of identity, as may be prescribed, and in the case of a subscriber being a body corporate, such particulars as may be prescribed; (f) the particulars of the persons mentioned in the articles as the first directors of the company, their names, including surnames or family names, the Director Identification Number, residential address, nationality and such other particulars including proof of identity as may be prescribed; and (g) the particulars of the interests of the persons mentioned in the articles as the first directors of the company in other firms or bodies corporate along with their consent to act as directors of the company in such form and manner as may be prescribed. (2) The Registrar on the basis of documents and information filed under sub-section (1) shall register all the documents and information referred to in that subsection in the register and issue a certificate of incorporation in the prescribed form to the effect that the proposed company is incorporated under this Act. (3) On and from the date mentioned in the certificate of incorporation issued under sub-section (2), the Registrar shall allot to the company a corporate identity number, which shall be a distinct identity for the company and which shall also be included in the certificate. (4) The company shall maintain and preserve at its registered office copies of all documents and information as originally filed under sub-section (1) till its dissolution under this Act. (5) If any person furnishes any false or incorrect particulars of any information or suppresses any material information, of which he is aware in any of the documents filed with the Registrar in relation to the registration of a company, he shall be liable for action under section 447. (6) Without prejudice to the provisions of sub-section (5) where, at any time after the incorporation of a company, it is proved that the company has been got incorporated by furnishing any false or incorrect information or representation or by suppressing any material fact or information in any of the documents or declaration filed or made for incorporating such company, or by any fraudulent action, the promoters, the persons named as the first directors of the company and the persons making declaration under clause (b) of subsection (1) shall each be liable for action under section 447. (7) Without prejudice to the provisions of sub-section (6), where a company has been got incorporated by furnishing any false or incorrect information or representation or by suppressing any material fact or information in any of the documents or declaration filed or made for incorporating such-- company or by any fraudulent action, the Tribunal may, on an application made to it, on being satisfied that the situation so warrants, (a) pass such orders, as it may think fit, for regulation of the management of the company including changes, if any, in its memorandum and articles, in public interest or in the interest of the company and its members and creditors; or (b) direct that liability of the members shall be unlimited; or (c) direct removal of the name of the company from the register of companies; or (d) pass an order for the winding up of the company; or (e) pass such other orders as it may deem fit: Provided that before making any order under this sub-section,-- (i) the company shall be given a reasonable opportunity of being heard in the matter; and (ii) the Tribunal shall take into consideration the transactions entered into by the company, including the obligations, if any, contracted or payment of any liability. 1. Subs. by Act 1 of 2018, s. 5, for "an affidavit" (w.e.f. 27-7-2018). Rules YearDescriptionHindi DescriptionFiles(Eng)Files(Hindi)30-03-2014 Chapter II - The Companies (Incorporation) Rules, 2014 01-05-2015 The Companies (Incorporation) Amendment Rules, 2015 29-05-2015 Companies (Incorporation) Second Amendment Rules, 2015	What records must a company maintain at its registered office according to The Companies Act, 2013, Section 7(4)?
  In this Act, unless the context otherwise requires,-- (1) "actionable claim" shall have the same meaning as assigned to it in section 3 of the Transfer of Property Act, 1882 (4 of 1882); (2) "address of delivery" means the address of the recipient of goods or services or both indicated on the tax invoice issued by a registered person for delivery of such goods or services or both; (3) "address on record" means the address of the recipient as available in the records of the supplier; (4) "adjudicating authority" means any authority, appointed or authorised to pass any order or decision under this Act, but does not include the 1[Central Board of Indirect Taxes and Customs], the Revisional Authority, the Authority for Advance Ruling, the Appellate Authority for Advance Ruling, 2[The National Appellate Authority for Advance Ruling] 3[the Appellate Authority, the Appellate Tribunal and the Authority referred to in sub-section (2) of section 171]; (5) "agent" means a person, including a factor, broker, commission agent, arhatia, del credere agent, an auctioneer or any other mercantile agent, by whatever name called, who carries on the business of supply or receipt of goods or services or both on behalf of another; (6) "aggregate turnover" means the aggregate value of all taxable supplies (excluding the value of inward supplies on which tax is payable by a person on reverse charge basis), exempt supplies, exports of goods or services or both and inter-State supplies of persons having the same Permanent Account Number, to be computed on all India basis but excludes central tax, State tax, Union territory tax, integrated tax and cess; (7) "agriculturist" means an individual or a Hindu Undivided Family who undertakes cultivation of land-- (a) by own labour, or (b) by the labour of family, or (c) by servants on wages payable in cash or kind or by hired labour under personal supervision or the personal supervision of any member of the family; (8) "Appellate Authority" means an authority appointed or authorised to hear appeals as referred to in section 107; (9) "Appellate Tribunal" means the Goods and Services Tax Appellate Tribunal constituted under section 109; (10) "appointed day" means the date on which the provisions of this Act shall come into force; (11) "assessment" means determination of tax liability under this Act and includes self assessment, re-assessment, provisional assessment, summary assessment and best judgment assessment; (12) "associated enterprises" shall have the same meaning as assigned to it in section 92A of the Income-tax Act, 1961 (43 of 1961); (13) "audit means" the examination of records, returns and other documents maintained or furnished by the registered person under this Act or the rules made thereunder or under any other law for the time being in force to verify the correctness of turnover declared, taxes paid, refund claimed and input tax credit availed, and to assess his compliance with the provisions of this Act or the rules made thereunder; (14) "authorised bank" shall mean a bank or a branch of a bank authorised by the Government to collect the tax or any other amount payable under this Act; (15) "authorised representative" means the representative as referred to in section 116; (16) "Board" means the 4[Central Board of Indirect Taxes and Customs] constituted under the Central Boards of Revenue Act, 1963 (54 of 1963); (17) "business" includes-- (a) any trade, commerce, manufacture, profession, vocation, adventure, wager or any other similar activity, whether or not it is for a pecuniary benefit; (b) any activity or transaction in connection with or incidental or ancillary to sub-clause (a); (c) any activity or transaction in the nature of sub-clause (a), whether or not there is volume, frequency, continuity or regularity of such transaction; (d) supply or acquisition of goods including capital goods and services in connection with commencement or closure of business; (e) provision by a club, association, society, or any such body (for a subscription or any other consideration) of the facilities or benefits to its members; (f) admission, for a consideration, of persons to any premises; (g) services supplied by a person as the holder of an office which has been accepted by him in the course or furtherance of his trade, profession or vocation; 5 [(h) activities of a race club including by way of totalisator or a license to book maker or activities of a licensed book maker in such club; and] (i) any activity or transaction undertaken by the Central Government, a State Government or any local authority in which they are engaged as public authorities; 6 * * * * * . (19) "capital goods" means goods, the value of which is capitalised in the books of account of the person claiming the input tax credit and which are used or intended to be used in the course or furtherance of business; (20) "casual taxable person" means a person who occasionally undertakes transactions involving supply of goods or services or both in the course or furtherance of business, whether as principal, agent or in any other capacity, in a State or a Union territory where he has no fixed place of business; (21) "central tax" means the central goods and services tax levied under section 9; (22) "cess" shall have the same meaning as assigned to it in the Goods and Services Tax (Compensation to States) Act; (23) "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); (24) "Commissioner" means the Commissioner of central tax and includes the Principal Commissioner of central tax appointed under section 3 and the Commissioner of integrated tax appointed under the Integrated Goods and Services Tax Act; (25) "Commissioner in the Board" means the Commissioner referred to in section 168; (26) "common portal" means the common goods and services tax electronic portal referred to in section 146; (27) "common working days" in respect of a State or Union territory shall mean such days in succession which are not declared as gazetted holidays by the Central Government or the concerned State or Union territory Government; (28) "company secretary" means a company secretary as defined in clause (c) of sub-section (1) of section 2 of the Company Secretaries Act, 1980 (56 of 1980); (29) "competent authority" means such authority as may be notified by the Government; (30) "composite supply" means a supply made by a taxable person to a recipient consisting of two or more taxable supplies of goods or services or both, or any combination thereof, which are naturally bundled and supplied in conjunction with each other in the ordinary course of business, one of which is a principal supply; Illustration.--- Where goods are packed and transported with insurance, the supply of goods, packing materials, transport and insurance is a composite supply and supply of goods is a principal supply; (31) "consideration" in relation to the supply of goods or services or both includes (a) any payment made or to be made, whether in money or otherwise, in respect of, in response to, or for the inducement of, the supply of goods or services or both, whether by the recipient or by any other person but shall not include any subsidy given by the Central Government or a State Government; (b) the monetary value of any act or forbearance, in respect of, in response to, or for the inducement of, the supply of goods or services or both, whether by the recipient or by any other person but shall not include any subsidy given by the Central Government or a State Government: Provided that a deposit given in respect of the supply of goods or services or both shall not be considered as payment made for such supply unless the supplier applies such deposit as consideration for the said supply; (32) "continuous supply of goods" means a supply of goods which is provided, or agreed to be provided, continuously or on recurrent basis, under a contract, whether or not by means of a wire, cable, pipeline or other conduit, and for which the supplier invoices the recipient on a regular or periodic basis and includes supply of such goods as the Government may, subject to such conditions, as it may, by notification, specify; (33) "continuous supply of services" means a supply of services which is provided, or agreed to be provided, continuously or on recurrent basis, under a contract, for a period exceeding three months with periodic payment obligations and includes supply of such services as the Government may, subject to such conditions, as it may, by notification, specify; (34) "conveyance" includes a vessel, an aircraft and a vehicle; (35) cost accountant" means a cost accountant as defined in 7 [clause (b)] of sub-section (1) of section 2 of the Cost and Works Accountants Act, 1959 (23 of 1959); (36) "Council" means the Goods and Services Tax Council established under article 279A of the Constitution; (37) "credit note" means a document issued by a registered person under sub-section (1) of section 34; (38) "debit note" means a document issued by a registered person under sub-section (3) of section 34; (39) "deemed exports" means such supplies of goods as may be notified under section 147; (40) "designated authority" means such authority as may be notified by the Board; (41) "document includes written or printed record of any sort and electronic record as defined in clause (t) of section 2 of the Information Technology Act, 2000 (21 of 2000); (42) "drawback" in relation to any goods manufactured in India and exported, means the rebate of duty, tax or cess chargeable on any imported inputs or on any domestic inputs or input services used in the manufacture of such goods; (43) "electronic cash ledger" means the electronic cash ledger referred to in sub-section (1) of section 49; (44) "electronic commerce" means the supply of goods or services or both, including digital products over digital or electronic network; (45) "electronic commerce operator" means any person who owns, operates or manages digital or electronic facility or platform for electronic commerce; (46) "electronic credit ledger" means the electronic credit ledger referred to in sub-section (2) of section 49; (47) "exempt supply" means supply of any goods or services or both which attracts nil rate of tax or which may be wholly exempt from tax under section 11, or under section 6 of the Integrated Goods and Services Tax Act, and includes non-taxable supply; (48) "existing law" means any law, notification, order, rule or regulation relating to levy and collection of duty or tax on goods or services or both passed or made before the commencement of this Act by Parliament or any Authority or person having the power to make such law, notification, order, rule or regulation; (49) "family" means,-- (i) the spouse and children of the person; and (ii) the parents, grand-parents, brothers and sisters of the person if they are wholly or mainly dependent on the said person; (50) "fixed establishment" means a place (other than the registered place of business) which is characterised by a sufficient degree of permanence and suitable structure in terms of human and technical resources to supply services, or to receive and use services for its own needs; (51) "Fund" means the Consumer Welfare Fund established under section 57; (52) "goods" means every kind of movable property other than money and securities but includes actionable claim, growing crops, grass and things attached to or forming part of the land which are agreed to be severed before supply or under a contract of supply; (53) "Government" means the Central Government; (54) "Goods and Services Tax (Compensation to States) Act" means the Goods and Services Tax (Compensation to States) Act, 2017 (15 of 2017); (55) "goods and services tax practitioner" means any person who has been approved under section 48 to act as such practitioner; (56) "India" means the territory of India as referred to in article 1 of the Constitution, its territorial waters, seabed and sub-soil underlying such waters, continental shelf, exclusive economic zone or any other maritime zone as referred to in the Territorial Waters, Continental Shelf, Exclusive Economic Zone and other Maritime Zones Act, 1976 (80 of 1976), and the air space above its territory and territorial waters; (57) "Integrated Goods and Services Tax Act" means the Integrated Goods and Services Tax Act, 2017 (13 of 2017); (58) "integrated tax" means the integrated goods and services tax levied under the Integrated Goods and Services Tax Act; (59)"input" means any goods other than capital goods used or intended to be used by a supplier in the course or furtherance of business; (60) "input service" means any service used or intended to be used by a supplier in the course or furtherance of business; (61) "Input Service Distributor" means an office of the supplier of goods or services or both which receives tax invoices issued under section 31 towards the receipt of input services and issues a prescribed document for the purposes of distributing the credit of central tax, State tax, integrated tax or Union territory tax paid on the said services to a supplier of taxable goods or services or both having the same Permanent Account Number as that of the said office; (62) "input tax in relation to a registered person," means the central tax, State tax, integrated tax or Union territory tax charged on any supply of goods or services or both made to him and includes (a) the integrated goods and services tax charged on import of goods; (b) the tax payable under the provisions of sub-sections (3) and (4) of section 9; (c) the tax payable under the provisions of sub-sections (3) and (4) of section 5 of the Integrated Goods and Services Tax Act; (d) the tax payable under the provisions of sub-sections (3) and (4) of section 9 of the respective State Goods and Services Tax Act; or (e) the tax payable under the provisions of sub-sections (3) and (4) of section 7 of the Union Territory Goods and Services Tax Act, but does not include the tax paid under the composition levy; (63) "input tax credit" means the credit of input tax; (64) "intra-State supply of goods" shall have the same meaning as assigned to it in section 8 of the Integrated Goods and Services Tax Act; (65) "intra-State supply of services" shall have the same meaning as assigned to it in section 8 of the Integrated Goods and Services Tax Act; (66) "invoice or tax invoice" means the tax invoice referred to in section 31; (67) "inward supply" in relation to a person, shall mean receipt of goods or services or both whether by purchase, acquisition or any other means with or without consideration; (68) "job work" means any treatment or process undertaken by a person on goods belonging to another registered person and the expression job worker shall be construed accordingly; (69) "local authority" means-- (a) a Panchayat as defined in clause (d) of article 243 of the Constitution; (b) a Municipality as defined in clause (e) of article 243P of the Constitution; (c) a Municipal Committee, a Zilla Parishad, a District Board, and any other authority legally entitled to, or entrusted by the Central Government or any State Government with the control or management of a municipal or local fund; (d) a Cantonment Board as defined in section 3 of the Cantonments Act, 2006 (41 of 2006); (e) a Regional Council or a District Council constituted under the Sixth Schedule to the Constitution; (f) a Development Board constituted under article 371 8[and article 371J] of the Constitution; or (g) a Regional Council constituted under article 371A of the Constitution; (70) "location of the recipient of services" means, (a) where a supply is received at a place of business for which the registration has been obtained, the location of such place of business; (b) where a supply is received at a place other than the place of business for which registration has been obtained (a fixed establishment elsewhere), the location of such fixed establishment; (c) where a supply is received at more than one establishment, whether the place of business or fixed establishment, the location of the establishment most directly concerned with the receipt of the supply; and (d) in absence of such places, the location of the usual place of residence of the recipient; (71) "location of the supplier of services" means,-- (a) where a supply is made from a place of business for which the registration has been obtained, the location of such place of business; (b) where a supply is made from a place other than the place of business for which registration has been obtained (a fixed establishment elsewhere), the location of such fixed establishment; (c) where a supply is made from more than one establishment, whether the place of business or fixed establishment, the location of the establishment most directly concerned with the provisions of the supply; and (d) in absence of such places, the location of the usual place of residence of the supplier; (72) "manufacture" means processing of raw material or inputs in any manner that results in emergence of a new product having a distinct name, character and use and the term manufacturer shall be construed accordingly; (73) "market value" shall mean the full amount which a recipient of a supply is required to pay in order to obtain the goods or services or both of like kind and quality at or about the same time and at the same commercial level where the recipient and the supplier are not related; (74) "mixed supply" means two or more individual supplies of goods or services, or any combination thereof, made in conjunction with each other by a taxable person for a single price where such supply does not constitute a composite supply. Illustration.-- A supply of a package consisting of canned foods, sweets, chocolates, cakes, dry fruits, aerated drinks and fruit juices when supplied for a single price is a mixed supply. Each of these items can be supplied separately and is not dependent on any other. It shall not be a mixed supply if these items are supplied separately; (75) "money" means the Indian legal tender or any foreign currency, cheque, promissory note, bill of exchange, letter of credit, draft, pay order, traveller cheque, money order, postal or electronic remittance or any other instrument recognised by the Reserve Bank of India when used as a consideration to settle an obligation or exchange with Indian legal tender of another denomination but shall not include any currency that is held for its numismatic value; (76) "motor vehicle" shall have the same meaning as assigned to it in clause (28) of section 2 of the Motor Vehicles Act, 1988 (59 of 1988); (77) "non-resident taxable person" means any person who occasionally undertakes transactions involving supply of goods or services or both, whether as principal or agent or in any other capacity, but who has no fixed place of business or residence in India; (78) "non-taxable supply" means a supply of goods or services or both which is not leviable to tax under this Act or under the Integrated Goods and Services Tax Act; (79)" non-taxable territory" means the territory which is outside the taxable territory; (80) "notification" means a notification published in the Official Gazette and the expressions notify and notified shall be construed accordingly; 9[(80A) online gaming means offering of a game on the internet or an electronic network and includes online money gaming; (80B) online money gaming means online gaming in which players pay or deposit money or money's worth, including virtual digital assets, in the expectation of winning money or money's worth, including virtual digital assets, in any event including game, scheme, competition or any other activity or process, whether or not its outcome or performance is based on skill, chance or both and whether the same is permissible or otherwise under any other law for the time being in force;] (81) "other territory" includes territories other than those comprising in a State and those referred to in sub-clauses (a) to (e) of clause (114) ; (82) "output tax" in relation to a taxable person, means the tax chargeable under this Act on taxable supply of goods or services or both made by him or by his agent but excludes tax payable by him on reverse charge basis; (83) "outward supply" in relation to a taxable person, means supply of goods or services or both, whether by sale, transfer, barter, exchange, licence, rental, lease or disposal or any other mode, made or agreed to be made by such person in the course or furtherance of business; (84) "person" includes-- (a) an individual; (b) a Hindu Undivided Family; (c) a company; (d) a firm; (e) a Limited Liability Partnership; (f) an association of persons or a body of individuals, whether incorporated or not, in India or outside India; (g) any corporation established by or under any Central Act, State Act or Provincial Act or a Government company as defined in clause (45) of section 2 of the Companies Act, 2013 (18 of 2013); (h) any body corporate incorporated by or under the laws of a country outside India; (i) a co-operative society registered under any law relating to co-operative societies; (j) a local authority; (k) Central Government or a State Government; (l) society as defined under the Societies Registration Act, 1860 (21 of 1860); (m) trust; and (n) every artificial juridical person, not falling within any of the above; (85) place of business includes-- (a) a place from where the business is ordinarily carried on, and includes a warehouse, a godown or any other place where a taxable person stores his goods, supplies or receives goods or services or both; or (b) a place where a taxable person maintains his books of account; or (c) a place where a taxable person is engaged in business through an agent, by whatever name called; (86) "place of supply" means the place of supply as referred to in Chapter V of the Integrated Goods and Services Tax Act; (87) "prescribed" means prescribed by rules made under this Act on the recommendations of the Council; (88) "principal" means a person on whose behalf an agent carries on the business of supply or receipt of goods or services or both; (89) "principal place of business" means the place of business specified as the principal place of business in the certificate of registration; (90) "principal supply" means the supply of goods or services which constitutes the predominant element of a composite supply and to which any other supply forming part of that composite supply is ancillary; (91) "proper officer" in relation to any function to be performed under this Act, means the Commissioner or the officer of the central tax who is assigned that function by the Commissioner in the Board; (92) "quarter" shall mean a period comprising three consecutive calendar months, ending on the last day of March, June, September and December of a calendar year; (93) "recipient" of supply of goods or services or both, means-- (a) where a consideration is payable for the supply of goods or services or both, the person who is liable to pay that consideration; (b) where no consideration is payable for the supply of goods, the person to whom the goods are delivered or made available, or to whom possession or use of the goods is given or made available; and (c) where no consideration is payable for the supply of a service, the person to whom the service is rendered, and any reference to a person to whom a supply is made shall be construed as a reference to the recipient of the supply and shall include an agent acting as such on behalf of the recipient in relation to the goods or services or both supplied; (94) "registered person" means a person who is registered under section 25 but does not include a person having a Unique Identity Number; (95) "regulations" means the regulations made by the Board under this Act on the recommendations of the Council; (96) "removal" in relation to goods, means-- (a) despatch of the goods for delivery by the supplier thereof or by any other person acting on behalf of such supplier; or (b) collection of the goods by the recipient thereof or by any other person acting on behalf of such recipient; (97) "return" means any return prescribed or otherwise required to be furnished by or under this Act or the rules made thereunder; (98) "reverse charge" means the liability to pay tax by the recipient of supply of goods or services or both instead of the supplier of such goods or services or both under sub-section (3) or sub-section (4) of section 9, or under sub-section (3) or sub-section (4) of section 5 of the Integrated Goods and Services Tax Act; (99) "Revisional Authority" means an authority appointed or authorised for revision of decision or orders as referred to in section 108; (100) "Schedule" means a Schedule appended to this Act; (101) "securities" shall have the same meaning as assigned to it in clause (h) of section 2 of the Securities Contracts (Regulation) Act, 1956 (42 of 1956); (102) "services" means anything other than goods, money and securities but includes activities relating to the use of money or its conversion by cash or by any other mode, from one form, currency or denomination, to another form, currency or denomination for which a separate consideration is charged; 10[Explanation. For the removal of doubts, it is hereby clarified that the expression services includes facilitating or arranging transactions in securities;] 11[(102A) specified actionable claim means the actionable claim involved in or by way of (i) betting; (ii) casinos; (iii) gambling; (iv) horse racing; (v) lottery; or (vi) online money gaming;ic9653 (103) State includes a Union territory with Legislature; (104) "State tax" means the tax levied under any State Goods and Services Tax Act; (105) "supplier" in relation to any goods or services or both, shall mean the person supplying the said goods or services or both and shall include an agent acting as such on behalf of such supplier in relation to the goods or services or both supplied; 12[Provided that a person who organises or arranges, directly or indirectly, supply of specified actionable claims, including a person who owns, operates or manages digital or electronic platform for such supply, shall be deemed to be a supplier of such actionable claims, whether such actionable claims are supplied by him or through him and whether consideration in money or money's worth, including virtual digital assets, for supply of such actionable claims is paid or conveyed to him or through him or placed at his disposal in any manner, and all the provisions of this Act shall apply to such supplier of specified actionable claims, as if he is the supplier liable to pay the tax in relation to the supply of such actionable claims;] (106) "tax period" means the period for which the return is required to be furnished; (107) "taxable person" means a person who is registered or liable to be registered under section 22 or section 24; (108) "taxable supply" means a supply of goods or services or both which is leviable to tax under this Act; (109) "taxable territory" means the territory to which the provisions of this Act apply; (110) "telecommunication service" means service of any description (including electronic mail, voice mail, data services, audio text services, video text services, radio paging and cellular mobile telephone services) which is made available to users by means of any transmission or reception of signs, signals, writing, images and sounds or intelligence of any nature, by wire, radio, visual or other electromagnetic means; (111) "the State Goods and Services Tax Act" means the respective State Goods and Services Tax Act, 2017; (112) "turnover in State or turnover in Union territory" means the aggregate value of all taxable supplies (excluding the value of inward supplies on which tax is payable by a person on reverse charge basis) and exempt supplies made within a State or Union territory by a taxable person, exports of goods or services or both and inter-State supplies of goods or services or both made from the State or Union territory by the said taxable person but excludes central tax, State tax, Union territory tax, integrated tax and cess; (113) "usual place of residence" means-- (a) in case of an individual, the place where he ordinarily resides; (b) in other cases, the place where the person is incorporated or otherwise legally constituted; (114) "Union territory" means the territory of-- (a) the Andaman and Nicobar Islands; (b) Lakshadweep; 13[(c) Dadra and Nagar Haveli and Daman and Diu; (d) Ladakh;] (e) Chandigarh; and 14*** 15[(ea) Ladakh; and; and] (f) other territory. Explanation-- For the purposes of this Act, each of the territories specified in sub-clauses (a) to (f) shall be considered to be a separate Union territory; (115) "Union territory tax" means the Union territory goods and services tax levied under the Union Territory Goods and Services Tax Act; (116) "Union Territory Goods and Services Tax Act" means the Union Territory Goods and Services Tax Act, 2017 (14 of 2017.); (117) "valid return" means a return furnished under sub-section (1) of section 39 on which selfassessed tax has been paid in full; 16[(117A) virtual digital asset shall have the same meaning as assigned to it in clause (47A) of Section 2 of the Income Tax Act, 1961 (43 of 1961);] (118) "voucher" means an instrument where there is an obligation to accept it as consideration or part consideration for a supply of goods or services or both and where the goods or services or both to be supplied or the identities of their potential suppliers are either indicated on the instrument itself or in related documentation, including the terms and conditions of use of such instrument; (119) "works contract" means a contract for building, construction, fabrication, completion, erection, installation, fitting out, improvement, modification, repair, maintenance, renovation, alteration or commissioning of any immovable property wherein transfer of property in goods (whether as goods or in some other form) is involved in the execution of such contract; (120) words and expressions used and not defined in this Act but defined in the Integrated Goods and Services Tax Act, the Union Territory Goods and Services Tax Act and the Goods and Services Tax (Compensation to States) Act shall have the same meaning as assigned to them in those Acts; 18* * * * * . 1. Subs. by Act 31 of 2018, s. 2, for Central Board of Excise and Customs (w.e.f. 1-2-2019). 2. Ins. by Act 23 of 2019, s. 92 (w.e.f. 1-1-2020). 3. Subs. by Act 31 of 2018, s. 2, for "the Appellate Authority and the Appellate Tribunal" (w.e.f. 1-2-2019). 4. Subs. by Act 31 of 2018, s. 2, for Central Board of Excise and Customs (w.e.f. 1-2-2019). 5. Subs. by s. 2, ibid., for clause (h) (w.e.f. 1-2-2019). 6. Clause (18) omitted by Act 31 of 2018, s. 2, (w.e.f. 1-2-2019). 7. Subs. by Act 31 of 2018 by s. 2, for clause (c) (w.e.f. 1-2-2019). 8. Ins. by Act 31 of 2018, s. 2 (w.e.f. 1-2-2019). 9. Ins. by Act 30 of 2023, s. 2 (w.e.f. 1-10-2023). 10.The Explanation ins. by Act 31 of 2018, s. 2 (w.e.f. 1-2-2019). 11 Ins. by Act 30 of 2023, s. 2 (w.e.f. 1-10-2023). 12. Ins. by Act 30 of 2023, s. 2 (w.e.f. 1-10-2023). 13.Subs. by Act 12 of 2020, s.118, for sub-clauses (c) and (d) (w.e.f. 30-6-2020). 14. The Word "and" omitted by the Jammu and Kashmir Reorganization (Adaptation of Central Laws) Order, 2020, vide notification No. S.O. 1123(E) dated (18-3-2020) and vide Union Territory of Ladakh Reorganisation (Adaptation of Central Laws) Order, 2020, notification No. S.O. 3774(E), dated (23-10-2020). 15. Vide notification No. S.O. 3912(E), dated 30th October, 2019, this Act is made applicable to the Union territory of Jammu and Kashmir and the Union territory of Ladakh. 16. Ins. by the Jammu and Kashmir Reorganization (Adaptation of Central Laws) Order, 2020, vide notification No. S.O. 1123(E) dated (18-3-2020) and vide Union Territory of Ladakh Reorganisation (Adaptation of Central Laws) Order, 2020, notification No. S.O. 3774(E), dated (23-10-2020). 17. ns. by Act 30 of 2023, s. 2 (w.e.f. 1-10-2023). 18.Clause (121) omitted by the Jammu and Kashmir Reorganization (Adaptation of Central Laws) Order, 2020, vide notification No. S.O. 1123(E) dated (18-3-2020) and vide Union Territory of Ladakh Reorganisation (Adaptation of Central Laws) Order, 2020, notification No. S.O. 3774(E), dated (23-10-2020).	How is a "valid return" defined in the Central Goods and Services Tax Act, 2017?

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512
      ],
      "matryoshka_weights": [
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: epoch
• gradient_accumulation_steps: 8
• 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: 8
• per_device_eval_batch_size: 8
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 8
• 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: 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
• eval_on_start: False
• use_liger_kernel: False
• eval_use_gather_object: False
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	dim_512_cosine_map@100	dim_768_cosine_map@100
0.0654	10	0.4082	-	-
0.1307	20	0.3974	-	-
0.1961	30	0.2993	-	-
0.2614	40	0.1891	-	-
0.3268	50	0.1704	-	-
0.3922	60	0.1915	-	-
0.4575	70	0.1223	-	-
0.5229	80	0.1679	-	-
0.5882	90	0.1619	-	-
0.6536	100	0.104	-	-
0.7190	110	0.0836	-	-
0.7843	120	0.1144	-	-
0.8497	130	0.1007	-	-
0.9150	140	0.0899	-	-
0.9804	150	0.1287	-	-
1.0	153	-	0.1586	0.1616
1.0458	160	0.0974	-	-
1.1111	170	0.0644	-	-
1.1765	180	0.0798	-	-
1.2418	190	0.0421	-	-
1.3072	200	0.04	-	-
1.3725	210	0.0445	-	-
1.4379	220	0.0309	-	-
1.5033	230	0.0282	-	-
1.5686	240	0.0367	-	-
1.6340	250	0.038	-	-
1.6993	260	0.0282	-	-
1.7647	270	0.0272	-	-
1.8301	280	0.0234	-	-
1.8954	290	0.0248	-	-
1.9608	300	0.0383	-	-
2.0	306	-	0.1588	0.1618
2.0261	310	0.0235	-	-
2.0915	320	0.0222	-	-
2.1569	330	0.0184	-	-
2.2222	340	0.0225	-	-
2.2876	350	0.0136	-	-
2.3529	360	0.0121	-	-
2.4183	370	0.0142	-	-
2.4837	380	0.0101	-	-
2.5490	390	0.0106	-	-
2.6144	400	0.0156	-	-
2.6797	410	0.0121	-	-
2.7451	420	0.0118	-	-
2.8105	430	0.0106	-	-
2.8758	440	0.0098	-	-
2.9412	450	0.0107	-	-
3.0	459	-	0.1585	0.1607
3.0065	460	0.0158	-	-
3.0719	470	0.0111	-	-
3.1373	480	0.0088	-	-
3.2026	490	0.0101	-	-
3.2680	500	0.0066	-	-
3.3333	510	0.0064	-	-
3.3987	520	0.0087	-	-
3.4641	530	0.0064	-	-
3.5294	540	0.0078	-	-
3.5948	550	0.0089	-	-
3.6601	560	0.0075	-	-
3.7255	570	0.0071	-	-
3.7908	580	0.0079	-	-
3.8562	590	0.0068	-	-
3.9216	600	0.0083	-	-
3.9869	610	0.0114	-	-
4.0	612	-	0.1594	0.1624

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.11.5
• Sentence Transformers: 3.0.1
• Transformers: 4.45.0
• PyTorch: 2.3.0+cu121
• Accelerate: 0.34.2
• Datasets: 3.0.0
• Tokenizers: 0.20.0

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}
}