MugheesAwan11/bge-base-securiti-dataset-1-v20

SentenceTransformer based on BAAI/bge-base-en-v1.5

This is a sentence-transformers model finetuned from BAAI/bge-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: BAAI/bge-base-en-v1.5
• Maximum Sequence Length: 512 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': 512, 'do_lower_case': True}) with Transformer model: BertModel 
  (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})
  (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("MugheesAwan11/bge-base-securiti-dataset-1-v20")
# Run inference
sentences = [
    'Careers View Events Spotlight Talks IDC Names Securiti a Worldwide Leader in Data Privacy View Events Spotlight Talks Education Contact Us Schedule a Demo Products By Use Cases By Roles Data Command Center View Learn more Asset and Data Discovery Discover dark and native data assets Learn more Data Access Intelligence & Governance Identify which users have access to sensitive data and prevent unauthorized access Learn more Data Privacy Automation PrivacyCenter.Cloud | Data Mapping | DSR Automation | Assessment Automation | Vendor Assessment | Breach Management | Privacy Notice Learn more Sensitive Data Intelligence Discover & Classify Structured and Unstructured Data | People Data Graph Learn more Data Flow Intelligence & Governance Prevent sensitive data sprawl through real-, Press Releases View Careers View Events Spotlight Talks IDC Names Securiti a Worldwide Leader in Data Privacy View Events Spotlight Talks Education Contact Us Schedule a Demo Products By Use Cases By Roles Data Command Center View Learn more Asset and Data Discovery Discover dark and native data assets Learn more Data Access Intelligence & Governance Identify which users have access to sensitive data and prevent unauthorized access Learn more Data Privacy Automation PrivacyCenter.Cloud | Data Mapping | DSR Automation | Assessment Automation | Vendor Assessment | Breach Management | Privacy Notice Learn more Sensitive Data Intelligence Discover & Classify Structured and Unstructured Data | People Data Graph Learn more Data Flow Intelligence & Governance Prevent',
    'What is the purpose of the Data Command Center?',
    "What are IBM's future prospects and preparedness for new business opportunities?",
]
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	0.3485
cosine_accuracy@3	0.5856
cosine_accuracy@5	0.6701
cosine_accuracy@10	0.7567
cosine_precision@1	0.3485
cosine_precision@3	0.1952
cosine_precision@5	0.134
cosine_precision@10	0.0757
cosine_recall@1	0.3485
cosine_recall@3	0.5856
cosine_recall@5	0.6701
cosine_recall@10	0.7567
cosine_ndcg@10	0.5507
cosine_mrr@10	0.4849
cosine_map@100	0.4942

Information Retrieval

• Dataset: dim_512
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.3464
cosine_accuracy@3	0.5938
cosine_accuracy@5	0.668
cosine_accuracy@10	0.7567
cosine_precision@1	0.3464
cosine_precision@3	0.1979
cosine_precision@5	0.1336
cosine_precision@10	0.0757
cosine_recall@1	0.3464
cosine_recall@3	0.5938
cosine_recall@5	0.668
cosine_recall@10	0.7567
cosine_ndcg@10	0.5518
cosine_mrr@10	0.486
cosine_map@100	0.4956

Information Retrieval

• Dataset: dim_256
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.3423
cosine_accuracy@3	0.567
cosine_accuracy@5	0.6619
cosine_accuracy@10	0.7485
cosine_precision@1	0.3423
cosine_precision@3	0.189
cosine_precision@5	0.1324
cosine_precision@10	0.0748
cosine_recall@1	0.3423
cosine_recall@3	0.567
cosine_recall@5	0.6619
cosine_recall@10	0.7485
cosine_ndcg@10	0.5413
cosine_mrr@10	0.4753
cosine_map@100	0.4846

Training Details

Training Dataset

Unnamed Dataset

• Size: 494 training samples
• Columns: positive and anchor
• Approximate statistics based on the first 1000 samples:

  	positive	anchor
  type	string	string
  details	min: 18 tokens mean: 223.56 tokens max: 414 tokens	min: 10 tokens mean: 21.87 tokens max: 102 tokens

• Samples:

  positive	anchor
  ### Denmark #### Denmark Effective Date : May 25, 2018 Region : EMEA (Europe, Middle East, Africa) Similar to other EU countries, Denmark has enacted a data protection act for the purpose of implementing the GDPR in the country. The Danish Data Protection Act (Act No. 502 of 23 May 2018) was enacted for the protection of natural persons with respect to personal data processing and to regulate the free movement of personal data. The Act replaced the previous Danish Act on Processing of Personal Data (Act no. 429 of 31/05/2000). Under the new Act, the Danish Data Protection Authority (Datatilsynet) oversees all aspects related to the supervision and enforcement of the Data Protection Act and the GDPR within the country as well as representing Denmark in the European Data Protection Board. ### Finland #### Finland Effective Date : January 1, 2019 Region : EMEA (Europe	What is the role of the Danish Data Protection Authority in Denmark's implementation of the GDPR?
  CPRA compliance involves adhering to the requirements outlined in the California Privacy Rights Act (CPRA) to protect consumer privacy and data rights. ## Join Our Newsletter Get all the latest information, law updates and more delivered to your inbox ### Share Copy 91 ### More Stories that May Interest You View More September 13, 2023 ## Kuwait's DPPR Kuwait didn’t have any data protection law until the Communication and Information Technology Regulatory Authority (CITRA) introduced the Data Privacy Protection Regulation (DPPR). The... View More September 11, 2023 ## Indonesia’s Protection of Personal Data Law: Explained In January 2020, Indonesia joined the burgeoning list of countries with their own data protection regulations. Provisions for data protection had existed within various... View More August 31, 2023 ##	Why is it important to comply with CPRA requirements and how does it protect data rights?
  Data Access Intelligence & Governance Identify which users have access to sensitive data and prevent unauthorized access Learn more Data Privacy Automation PrivacyCenter.Cloud	Data Mapping

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256
      ],
      "matryoshka_weights": [
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: epoch
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 16
• 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: 32
• 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
• 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
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	dim_256_cosine_map@100	dim_512_cosine_map@100	dim_768_cosine_map@100
0.625	10	3.7981	-	-	-
1.0	16	-	0.4653	0.4819	0.4810
1.25	20	2.2066	-	-	-
1.875	30	1.668	-	-	-
2.0	32	-	0.4837	0.4905	0.4933
2.5	40	0.9807	-	-	-
3.0	48	-	0.4846	0.4954	0.4949
3.125	50	1.0226	-	-	-
3.75	60	1.0564	-	-	-
4.0	64	-	0.4846	0.4956	0.4942

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.10.14
• Sentence Transformers: 3.0.1
• Transformers: 4.41.2
• PyTorch: 2.1.2+cu121
• Accelerate: 0.31.0
• 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",
}

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}
}