This is a sentence-transformers model finetuned from jxm/cde-small-v1 on the csv dataset. It maps sentences & paragraphs to a None-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: jxm/cde-small-v1
Maximum Sequence Length: None tokens
Output Dimensionality: None 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({}) with Transformer model: DatasetTransformer 
)

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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("jebish7/cde-small-v1_MNR_3")
# Run inference
sentences = [
    'What are the common scenarios or instances where assets and liabilities are not covered by the bases of accounting in Rule 5.3.2, and how should an Insurer address these in their reporting?',
    'DocumentID: 12 | PassageID: 5.3.1.Guidance | Passage: \nThe exceptions provided in this Chapter relate to the following:\na.\tspecific Rules in respect of certain assets and liabilities, intended to achieve a regulatory objective not achieved by application of either or both of the bases of accounting set out in Rule \u200e5.3.2;\nb.\tassets and liabilities that are not dealt with in either or both of the bases of accounting set out in Rule \u200e5.3.2; and\nc.\tthe overriding power of the Regulator, set out in Rule \u200e5.1.6, to require an Insurer to adopt a particular measurement for a specific asset or liability.',
    'DocumentID: 1 | PassageID: 14.4.1.Guidance.1. | Passage: Relevant Persons are reminded that in accordance with Federal AML Legislation, Relevant Persons or any of their Employees must not tip off any Person, that is, inform any Person that he is being scrutinised, or investigated by any other competent authority, for possible involvement in suspicious Transactions or activity related to money laundering or terrorist financing.',
]
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: 34.95 tokens
max: 68 tokens
min: 35 tokens
mean: 132.0 tokens
max: 512 tokens

Samples:

anchor	positive
`If a financial institution offers Money Remittance as one of its services, under what circumstances is it deemed to be holding Relevant Money and therefore subject to regulatory compliance (a)?`	`DocumentID: 13`
`What are the consequences for a Recognised Body or Authorised Person if they fail to comply with ADGM's requirements regarding severance payments?`	`DocumentID: 7`
`If a Public Fund is structured as an Investment Trust, to whom should the Fund Manager report the review findings regarding delegated Regulated Activities or outsourced functions?`	`DocumentID: 6`

Loss: MultipleNegativesRankingLoss with these parameters:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim" 4}

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 16
learning_rate: 2e-05
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: 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: 3
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
0.1082	100	1.9962
0.2165	200	1.1626
0.3247	300	0.9907
0.4329	400	0.8196
0.5411	500	0.8082
0.6494	600	0.6944
0.7576	700	0.6559
0.8658	800	0.6242
0.9740	900	0.6299
1.0823	1000	0.6051
1.1905	1100	0.567
1.2987	1200	0.4679
1.4069	1300	0.3443
1.5152	1400	0.3356
1.6234	1500	0.2958
1.7316	1600	0.254
1.8398	1700	0.2694
1.9481	1800	0.2497
2.0563	1900	0.2671
2.1645	2000	0.2558
2.2727	2100	0.1943
2.3810	2200	0.1242
2.4892	2300	0.116
2.5974	2400	0.1081
2.7056	2500	0.1056
2.8139	2600	0.107
2.9221	2700	0.1154

Framework Versions

Python: 3.10.14
Sentence Transformers: 3.1.1
Transformers: 4.45.2
PyTorch: 2.4.0
Accelerate: 0.34.2
Datasets: 3.0.1
Tokenizers: 0.20.0

Citation

BibTeX

Sentence Transformers

bibtex

@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",
}

MultipleNegativesRankingLoss

bibtex

@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}
}

	anchor	positive
type	string	string
details	min: 16 tokens mean: 34.95 tokens max: 68 tokens	min: 35 tokens mean: 132.0 tokens max: 512 tokens