metadata

base_model: sentence-transformers/all-mpnet-base-v2
datasets: []
language: []
library_name: sentence-transformers
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:80
  - loss:MultipleNegativesRankingLoss
widget:
  - source_sentence: What is the short title of this Act?
    sentences:
      - The Banking (Amendment) Act, No. 24 of 2024
      - >-
        The Central Bank can assess whether granting a license is in the
        interest of the banking system and national economy
      - >-
        They must conduct offshore banking business only with non-residents and
        certain residents as determined by the Central Bank
  - source_sentence: >-
      What new power does the Central Bank have regarding excess accommodations
      to related parties?
    sentences:
      - Accommodation and all financial investments
      - >-
        The Central Bank can assess a bank's track record for operating with
        good governance and integrity
      - >-
        The Central Bank may require the bank to deduct such excess from
        regulatory capital calculations
  - source_sentence: >-
      What new requirement is there for licensed commercial banks regarding the
      disclosure of audit information?
    sentences:
      - >-
        Auditors must immediately report findings that could materially affect
        the bank's safety and soundness
      - >-
        They must prepare financial statements that represent a true and fair
        view of the bank's financial position
      - >-
        To the Governor, Deputy Governor, Assistant Governor, Director of Bank
        Supervision, or other high-ranking officers
  - source_sentence: >-
      Who approves the appointment, election, or nomination of directors for
      licensed commercial banks?
    sentences:
      - The Central Bank
      - >-
        The Central Bank can designate foreign currencies for offshore banking
        business
      - >-
        The Central Bank can request information or documents from any person
        other than a licensed bank
  - source_sentence: >-
      What new power does the Central Bank have regarding the number of
      directors on a bank's board?
    sentences:
      - Within fifteen days of becoming aware of such facts
      - Transparency in ownership structure and beneficial ownership
      - >-
        The Central Bank can determine the number of board members, which shall
        not be less than seven

SentenceTransformer based on sentence-transformers/all-mpnet-base-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-mpnet-base-v2. 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: sentence-transformers/all-mpnet-base-v2
Maximum Sequence Length: 384 tokens
Output Dimensionality: 768 tokens
Similarity Function: Cosine Similarity

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': 384, 'do_lower_case': False}) with Transformer model: MPNetModel 
  (1): Pooling({'word_embedding_dimension': 768, '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): Normalize()
)

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("sentence_transformers_model_id")
# Run inference
sentences = [
    "What new power does the Central Bank have regarding the number of directors on a bank's board?",
    'The Central Bank can determine the number of board members, which shall not be less than seven',
    'Within fifteen days of becoming aware of such facts',
]
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]

Training Details

Training Dataset

Unnamed Dataset

Size: 80 training samples
Columns: query and answer
Approximate statistics based on the first 1000 samples:
query answer
type string string
details
min: 8 tokens
mean: 16.82 tokens
max: 22 tokens

min: 3 tokens
mean: 16.85 tokens
max: 28 tokens

	query	answer
type	string	string
details	min: 8 tokens mean: 16.82 tokens max: 22 tokens	min: 3 tokens mean: 16.85 tokens max: 28 tokens

Samples:

query	answer
`What new power does the Central Bank have regarding corrective action for misleading disclosures?`	`The Central Bank can require banks to publish corrected disclosure statements`
`What new requirement is there for licensed commercial banks regarding corporate governance?`	`The Board of Directors is ultimately responsible for ensuring compliance with laws and sound banking practices`
`Who is responsible for overseeing the management of a licensed commercial bank?`	`The Board of Directors`

Loss: MultipleNegativesRankingLoss with these parameters:

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

Evaluation Dataset

Unnamed Dataset

Size: 20 evaluation samples
Columns: query and answer
Approximate statistics based on the first 1000 samples:
query answer
type string string
details
min: 11 tokens
mean: 17.4 tokens
max: 23 tokens

min: 5 tokens
mean: 17.2 tokens
max: 29 tokens

	query	answer
type	string	string
details	min: 11 tokens mean: 17.4 tokens max: 23 tokens	min: 5 tokens mean: 17.2 tokens max: 29 tokens

Samples:

query	answer
`What new power does the Central Bank have regarding the assessment of bank license applications?`	`The Central Bank can assess whether granting a license is in the interest of the banking system and national economy`
`What new definition is provided for 'unlawful activity' in the Act?`	`It has the same meaning as in the Prevention of Money Laundering Act, No. 5 of 2006`
`What new requirement is there for licensed commercial banks regarding offshore banking?`	`They must conduct offshore banking business only with non-residents and certain residents as determined by the Central Bank`

Loss: MultipleNegativesRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_eval_batch_size: 16
learning_rate: 3e-05
warmup_ratio: 0.1
fp16: True
load_best_model_at_end: True
ddp_find_unused_parameters: False

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: 16
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: 3e-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: True
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: 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
optim_args: None
adafactor: False
group_by_length: False
length_column_name: length
ddp_find_unused_parameters: False
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
eval_use_gather_object: False
batch_sampler: batch_sampler
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	loss
0.2	2	1.1686	-
0.4	4	1.0061	-
0.6	6	0.8377	-
0.8	8	0.8121	-
1.0	10	0.267	0.8161
1.2	12	0.3282	-
1.4	14	0.1546	-
1.6	16	0.3546	-
1.8	18	0.0544	-
2.0	20	0.0271	0.6723
2.2	22	0.1496	-
2.4	24	0.0569	-
2.6	26	0.1682	-
2.8	28	0.0452	-
3.0	30	0.0261	0.6644

Framework Versions

Python: 3.10.14
Sentence Transformers: 3.0.1
Transformers: 4.43.1
PyTorch: 2.3.0
Accelerate: 0.32.0
Datasets: 2.19.1
Tokenizers: 0.19.1

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",
}

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}
}