metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:29545
  - loss:MultipleNegativesSymmetricRankingLoss
base_model: dunzhang/stella_en_400M_v5
widget:
  - source_sentence: >-
      In the context of the risk-based assessment of customers and business
      relationships, how should the overlap between customer risk assessment and
      CDD be managed to ensure both are completed effectively and in compliance
      with ADGM regulations?
    sentences:
      - >
        DocumentID: 36 | PassageID: D.7. | Passage: Principle 7 – Scenario
        analysis of climate-related financial risks. Where appropriate, relevant
        financial firms should develop and implement climate-related scenario
        analysis frameworks, including stress testing, in a manner commensurate
        with their size, complexity, risk profile and nature of activities.
      - >-
        DocumentID: 1 | PassageID: 7.Guidance.4. | Passage: The risk-based
        assessment of the customer and the proposed business relationship,
        Transaction or product required under this Chapter is required to be
        undertaken prior to the establishment of a business relationship with a
        customer. Because the risk rating assigned to a customer resulting from
        this assessment determines the level of CDD that must be undertaken for
        that customer, this process must be completed before the CDD is
        completed for the customer. The Regulator is aware that in practice
        there will often be some degree of overlap between the customer risk
        assessment and CDD. For example, a Relevant Person may undertake some
        aspects of CDD, such as identifying Beneficial Owners, when it performs
        a risk assessment of the customer. Conversely, a Relevant Person may
        also obtain relevant information as part of CDD which has an impact on
        its customer risk assessment. Where information obtained as part of CDD
        of a customer affects the risk rating of a customer, the change in risk
        rating should be reflected in the degree of CDD undertaken.
      - >-
        DocumentID: 1 | PassageID: 9.1.2.Guidance.4. | Passage: Where the
        legislative framework of a jurisdiction (such as secrecy or data
        protection legislation) prevents a Relevant Person from having access to
        CDD information upon request without delay as referred to in Rule
        ‎9.1.1(3)(b), the Relevant Person should undertake the relevant CDD
        itself and should not seek to rely on the relevant third party.
  - source_sentence: >-
      Can you clarify the responsibilities of the Governing Body of a Relevant
      Person in establishing and maintaining AML/TFS policies and procedures,
      and how these should be documented and reviewed?
    sentences:
      - >+
        DocumentID: 28 | PassageID: 193) | Passage: SUPERVISION BY LISTING
        AUTHORITY

        Complaints or allegations of non-compliance by Reporting Entities

        If, as a result of the enquiry, the Listing Authority forms the view
        that the information is accurate, is Inside Information, and is not
        within exemption from Disclosure provided by Rule 7.2.2, the Listing
        Authority will ask the Reporting Entity to make a Disclosure about the
        matter under Rule 7.2.1.  If the information should have been Disclosed
        earlier, the Listing Authority may issue an ‘aware letter’ (see
        paragraphs 187 to 189 above), or take other relevant action.

      - "DocumentID: 17 | PassageID: Part 13.165.(2) | Passage: The Regulator shall not approve a Non Abu Dhabi Global Market Clearing House unless it is satisfied—\n(a)\tthat the rules and practices of the body, together with the law of the country in which the body's head office is situated, provide adequate procedures for dealing with the default of persons party to contracts connected with the body; and\n(b)\tthat it is otherwise appropriate to approve the body;\ntogether being the “Relevant Requirements” for this Part."
      - "DocumentID: 1 | PassageID: 4.3.1 | Passage: A Relevant Person which is part of a Group must ensure that it:\n(a)\thas developed and implemented policies and procedures for the sharing of information between Group entities, including the sharing of information relating to CDD and money laundering risks;\n(b)\thas in place adequate safeguards on the confidentiality and use of information exchanged between Group entities, including consideration of relevant data protection legislation;\n(c)\tremains aware of the money laundering risks of the Group as a whole and of its exposure to the Group and takes active steps to mitigate such risks;\n(d)\tcontributes to a Group-wide risk assessment to identify and assess money laundering risks for the Group; and\n(e)\tprovides its Group-wide compliance, audit and AML/TFS functions with customer account and Transaction information from its Branches and Subsidiaries when necessary for AML/TFS purposes."
  - source_sentence: >-
      What specific accounting standards and practices are we required to follow
      when valuing positions in our Trading and Non-Trading Books to ensure
      compliance with ADGM regulations?
    sentences:
      - >
        DocumentID: 7 | PassageID: 8.10.1.(2).Guidance.3. | Passage: Each
        Authorised Person, Recognised Body and its Auditors is also required
        under Part 16 and section 193 of the FSMR respectively, to disclose to
        the Regulator any matter which may indicate a breach or likely breach
        of, or a failure or likely failure to comply with, Regulations or Rules.
        Each Authorised Person and Recognised Body is also required to establish
        and implement systems and procedures to enable its compliance and
        compliance by its Auditors with notification requirements.
      - "DocumentID: 18 | PassageID: 3.2 | Passage: Financial Services Permissions. VC Managers operating in ADGM require a Financial Services Permission (“FSP”) to undertake any Regulated Activity pertaining to VC Funds and/or co-investments by third parties in VC Funds. The Regulated Activities covered by the FSP will be dependent on the VC Managers’ investment strategy and business model.\n(a)\tManaging a Collective Investment Fund: this includes carrying out fund management activities in respect of a VC Fund.\n(b)\tAdvising on Investments or Credit : for VC Managers these activities will be restricted to activities related to co-investment alongside a VC Fund which the VC Manager manages, such as recommending that a client invest in an investee company alongside the VC Fund and on the strategy and structure required to make the investment.\n(c)\tArranging Deals in Investments: VC Managers may also wish to make arrangements to facilitate co-investments in the investee company.\nAuthorisation fees and supervision fees for a VC Manager are capped at USD 10,000 regardless of whether one or both of the additional Regulated Activities in b) and c) above in relation to co-investments are included in its FSP. The FSP will include restrictions appropriate to the business model of a VC Manager."
      - >
        DocumentID: 13 | PassageID: APP2.A2.1.1.(4) | Passage: An Authorised
        Person must value every position included in its Trading Book and the
        Non Trading Book in accordance with the relevant accounting standards
        and practices.
  - source_sentence: >-
      What documentation and information are we required to maintain to
      demonstrate compliance with the rules pertaining to the cooperation with
      auditors, especially in terms of providing access and not interfering with
      their duties?
    sentences:
      - "DocumentID: 6 | PassageID: PART 5.16.3.5 | Passage: Co-operation with auditors. A Fund Manager must take reasonable steps to ensure that it and its Employees:\n(a)\tprovide any information to its auditor that its auditor reasonably requires, or is entitled to receive as auditor;\n(b)\tgive the auditor right of access at all reasonable times to relevant records and information within its possession;\n(c)\tallow the auditor to make copies of any records or information referred to in ‎(b);\n(d)\tdo not interfere with the auditor's ability to discharge its duties;\n(e)\treport to the auditor any matter which may significantly affect the financial position of the Fund; and\n(f)\tprovide such other assistance as the auditor may reasonably request it to provide."
      - "DocumentID: 13 | PassageID: 4.3.1 | Passage: An Authorised Person must implement and maintain comprehensive Credit Risk management systems which:\n(a)\tare appropriate to the firm's type, scope, complexity and scale of operations;\n(b)\tare appropriate to the diversity of its operations, including geographical diversity;\n(c)\tenable the firm to effectively identify, assess, monitor and control Credit Risk and to ensure that adequate Capital Resources are available at all times to cover the risks assumed; and\n(d)\tensure effective implementation of the Credit Risk strategy and policy."
      - >-
        DocumentID: 3 | PassageID: 3.8.9 | Passage: The Authorised Person acting
        as the Investment Manager of an ADGM Green Portfolio must provide a copy
        of the attestation obtained for the purposes of Rule ‎3.8.6 to each
        Client with whom it has entered into a Discretionary Portfolio
        Management Agreement in respect of such ADGM Green Portfolio at least on
        an annual basis and upon request by the Client.
  - source_sentence: >-
      Could you provide examples of circumstances that, when changed, would
      necessitate the reevaluation of a customer's risk assessment and the
      application of updated CDD measures?
    sentences:
      - >-
        DocumentID: 13 | PassageID: 9.2.1.Guidance.1. | Passage: The Regulator
        expects that an Authorised Person's Liquidity Risk strategy will set out
        the approach that the Authorised Person will take to Liquidity Risk
        management, including various quantitative and qualitative targets. It
        should be communicated to all relevant functions and staff within the
        organisation and be set out in the Authorised Person's Liquidity Risk
        policy.
      - "DocumentID: 1 | PassageID: 8.1.2.(1) | Passage: A Relevant Person must also apply CDD measures to each existing customer under Rules ‎8.3.1, ‎8.4.1 or ‎8.5.1 as applicable:\n(a)\twith a frequency appropriate to the outcome of the risk-based approach taken in relation to each customer; and\n(b)\twhen the Relevant Person becomes aware that any circumstances relevant to its risk assessment for a customer have changed."
      - "DocumentID: 1 | PassageID: 8.1.1.Guidance.2. | Passage: The FIU has issued guides that require:\n(a)\ta DNFBP that is a dealer in precious metals or precious stones to obtain relevant identification documents, such as passport, emirates ID, trade licence, as applicable, and register the information via goAML for all cash transactions equal to or exceeding USD15,000 with individuals and all cash or wire transfer transactions equal to or exceeding USD15,000 with entities. The Regulator expects a dealer in any saleable item or a price equal to or greater than USD15,000 to also comply with this requirement;\n(b)\ta DNFBP that is a real estate agent to obtain relevant identification documents, such as passport, emirates ID, trade licence, as applicable, and register the information via goAML for all sales or purchases of Real Property where:\n(i)\tthe payment for the sale/purchase includes a total cash payment of USD15,000 or more whether in a single cash payment or multiple cash payments;\n(ii)\tthe payment for any part or all of the sale/purchase amount includes payment(s) using Virtual Assets;\n(iii)\tthe payment for any part or all of the sale/purchase amount includes funds that were converted from or to a Virtual Asset."
pipeline_tag: sentence-similarity
library_name: sentence-transformers

SentenceTransformer based on dunzhang/stella_en_400M_v5

This is a sentence-transformers model finetuned from dunzhang/stella_en_400M_v5 on the csv dataset. It maps sentences & paragraphs to a 1024-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: dunzhang/stella_en_400M_v5
Maximum Sequence Length: 512 tokens
Output Dimensionality: 1024 tokens
Similarity Function: Cosine Similarity
Training Dataset:
- csv

Model Sources

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

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False}) with Transformer model: NewModel 
  (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
  (2): Dense({'in_features': 1024, 'out_features': 1024, 'bias': True, 'activation_function': 'torch.nn.modules.linear.Identity'})
)

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("jebish7/stella-MNSR-3")
# Run inference
sentences = [
    "Could you provide examples of circumstances that, when changed, would necessitate the reevaluation of a customer's risk assessment and the application of updated CDD measures?",
    'DocumentID: 1 | PassageID: 8.1.2.(1) | Passage: A Relevant Person must also apply CDD measures to each existing customer under Rules \u200e8.3.1, \u200e8.4.1 or \u200e8.5.1 as applicable:\n(a)\twith a frequency appropriate to the outcome of the risk-based approach taken in relation to each customer; and\n(b)\twhen the Relevant Person becomes aware that any circumstances relevant to its risk assessment for a customer have changed.',
    "DocumentID: 13 | PassageID: 9.2.1.Guidance.1. | Passage: The Regulator expects that an Authorised Person's Liquidity Risk strategy will set out the approach that the Authorised Person will take to Liquidity Risk management, including various quantitative and qualitative targets. It should be communicated to all relevant functions and staff within the organisation and be set out in the Authorised Person's Liquidity Risk policy.",
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

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

Training Details

Training Dataset

csv

Dataset: csv
Size: 29,545 training samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 16 tokens
mean: 35.04 tokens
max: 64 tokens

min: 27 tokens
mean: 129.43 tokens
max: 512 tokens

	anchor	positive
type	string	string
details	min: 16 tokens mean: 35.04 tokens max: 64 tokens	min: 27 tokens mean: 129.43 tokens max: 512 tokens

Samples:

anchor	positive
`Could you outline the expected procedures for a Trade Repository to notify relevant authorities of any significant errors or omissions in previously submitted data?`	`DocumentID: 7`
`In the context of a non-binding MPO, how are commodities held by an Authorised Person treated for the purpose of determining the Commodities Risk Capital Requirement?`	`DocumentID: 9`
`Can the FSRA provide case studies or examples of best practices for RIEs operating MTFs or OTFs using spot commodities in line with the Spot Commodities Framework?`	`DocumentID: 34`

Loss: MultipleNegativesSymmetricRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 16
num_train_epochs: 1
warmup_ratio: 0.1
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: no
prediction_loss_only: True
per_device_train_batch_size: 16
per_device_eval_batch_size: 8
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 1
eval_accumulation_steps: None
torch_empty_cache_steps: None
learning_rate: 5e-05
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 1
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.1
warmup_steps: 0
log_level: passive
log_level_replica: warning
log_on_each_node: True
logging_nan_inf_filter: True
save_safetensors: True
save_on_each_node: False
save_only_model: False
restore_callback_states_from_checkpoint: False
no_cuda: False
use_cpu: False
use_mps_device: False
seed: 42
data_seed: None
jit_mode_eval: False
use_ipex: False
bf16: False
fp16: False
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: None
local_rank: 0
ddp_backend: None
tpu_num_cores: None
tpu_metrics_debug: False
debug: []
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_prefetch_factor: None
past_index: -1
disable_tqdm: False
remove_unused_columns: True
label_names: None
load_best_model_at_end: False
ignore_data_skip: False
fsdp: []
fsdp_min_num_params: 0
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
fsdp_transformer_layer_cls_to_wrap: None
accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
deepspeed: None
label_smoothing_factor: 0.0
optim: adamw_torch
optim_args: None
adafactor: False
group_by_length: False
length_column_name: length
ddp_find_unused_parameters: None
ddp_bucket_cap_mb: None
ddp_broadcast_buffers: False
dataloader_pin_memory: True
dataloader_persistent_workers: False
skip_memory_metrics: True
use_legacy_prediction_loop: False
push_to_hub: False
resume_from_checkpoint: None
hub_model_id: None
hub_strategy: every_save
hub_private_repo: 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
0.0541	100	0.4442
0.1083	200	0.4793
0.1624	300	0.4395
0.2166	400	0.4783
0.2707	500	0.4573
0.3249	600	0.4235
0.3790	700	0.4029
0.4331	800	0.3951
0.4873	900	0.438
0.5414	1000	0.364
0.5956	1100	0.3732
0.6497	1200	0.3932
0.7038	1300	0.3387
0.7580	1400	0.2956
0.8121	1500	0.3612
0.8663	1600	0.3333
0.9204	1700	0.2837
0.9746	1800	0.2785
0.0541	100	0.2263
0.1083	200	0.2085
0.1624	300	0.1638
0.2166	400	0.2085
0.2707	500	0.2442
0.3249	600	0.1965
0.3790	700	0.2548
0.4331	800	0.2504
0.4873	900	0.2358
0.5414	1000	0.2083
0.5956	1100	0.2117
0.6497	1200	0.248
0.7038	1300	0.221
0.7580	1400	0.1886
0.8121	1500	0.2653
0.8663	1600	0.2651
0.9204	1700	0.2349
0.9746	1800	0.2435
0.0541	100	0.143
0.1083	200	0.0701
0.1624	300	0.0675
0.2166	400	0.0977
0.2707	500	0.1157
0.3249	600	0.0823
0.3790	700	0.1022
0.4331	800	0.114
0.4873	900	0.0955
0.5414	1000	0.0905
0.5956	1100	0.0959
0.6497	1200	0.1308
0.7038	1300	0.1285
0.7580	1400	0.1006
0.8121	1500	0.1553
0.8663	1600	0.1769
0.9204	1700	0.1965
0.9746	1800	0.2271

Framework Versions

Python: 3.10.12
Sentence Transformers: 3.1.1
Transformers: 4.45.2
PyTorch: 2.5.1+cu121
Accelerate: 1.1.1
Datasets: 3.1.0
Tokenizers: 0.20.3

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",
}