jebish7/cde-small-v1_MNR_3

SentenceTransformer based on jxm/cde-small-v1

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.

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:

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

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

@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

@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}
}