bge-base-privacy-act-ft

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 retrieval.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: BAAI/bge-base-en-v1.5 <!-- at revision a5beb1e3e68b9ab74eb54cfd186867f64f240e1a -->
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Supported Modality: Text

<!-- - **Training Dataset:** Unknown --> <!-- - **Language:** Unknown --> <!-- - **License:** Unknown -->

Model Sources

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

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'transformer_task': 'feature-extraction', 'modality_config': {'text': {'method': 'forward', 'method_output_name': 'last_hidden_state'}}, 'module_output_name': 'token_embeddings', 'architecture': 'BertModel'})
  (1): Pooling({'embedding_dimension': 768, 'pooling_mode': 'cls', '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("greecehalf/bge-base-privacy-act-ft")
# Run inference
sentences = [
    'What is an infringement notice under the Privacy Act?',
    'Part VIB  Compliance and enforcement \nDivision 1A  Infringement notices \n \nSection 80UB \n \n334 Privacy Act 1988  \nCompilation No. 104  Compilation date: 04/06/2026 \n \nDivision 1A—Infringement notices \n80UB  Infringement notices \nProvisions subject to an infringement notice \n (1) The following provisions are subject to an infringement notice \nunder Part 5 of the Regulatory Powers Act: \n (a) subsections 13K(1) and (2) (civil penalty provision for which \ninfringement notices or compliance notices can be issued); \n (b) subsection 66(1) (failure to give information); \n (c) subsection 80UC(4) (failure to comply with a compliance \nnotice). \nNote: Part 5 of the Regulatory Powers Act creates a framework for using \ninfringement notices in relation to provisions. \nAmount to be stated in an infringement notice for listed \ncompanies—section 13K \n (1A) Despite subsection 104(2) of the Regulatory Powers Act, if an \ninfringement notice relates to only one alleged contravention of \nsubsection 13K(1) or (2), or 80UC(4), by a listed corporation \n(within the meaning of the Corporations Act 2001), the amount to \nbe stated in the notice for the purposes of paragraph 104(1)(f) of \nthe Regulatory Powers Act is 200 penalty units. \n (1B) Despite subsection 104(3) of the Regulatory Powers Act, if an \ninfringement notice relates to more than one alleged contravention \nof subsection 13K(1) or (2), or 80UC(4), by a listed corporation \n(within the meaning of the Corporations Act 2001), the amount to',
    'Part II  Interpretation \nDivision 1  General definitions \n \nSection 6 \n \n16 Privacy Act 1988  \nCompilation No. 104  Compilation date: 04/06/2026 \n \ncredit reporting complaint means a complaint about an act or \npractice that, if established, would be an interference with the \nprivacy of an individual because: \n (a) it breached the registered CR code; or \n (b) it breached a provision of Part IIIA. \ncredit reporting information about an individual means credit \ninformation, or CRB derived information, about the individual. \ncredit worthiness of an individual means the individual’s: \n (a) eligibility to be provided with consumer credit; or \n (b) history in relation to consumer credit; or \n (c) capacity to repay an amount of credit that relates to consumer \ncredit. \nde facto partner of an individual has the meaning given by the Acts \nInterpretation Act 1901. \ndefault information has the meaning given by section 6Q. \nDefence Department means the Department of State that deals \nwith defence and that is administered by the Minister administering \nsection 1 of the Defence Act 1903. \nDefence Force includes the Australian Defence Force Cadets. \nde-identified: personal information is de-identified if the \ninformation is no longer about an identifiable individual or an \nindividual who is reasonably identifiable. \nDepartment means an Agency within the meaning of the Public \nService Act 1999. \neligible data breach has the meaning given by Division 2 of \nPart IIIC.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.8246, 0.4729],
#         [0.8246, 1.0000, 0.3557],
#         [0.4729, 0.3557, 1.0000]])

<!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use *List how the model may foreseeably be misused and address what users ought not to do with the model.* --> <!-- ## Bias, Risks and Limitations *What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.* --> <!-- ### Recommendations *What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.* -->

Training Details

Training Dataset

Unnamed Dataset

• Size: 36 training samples

• Columns: <code>sentence_0</code> and <code>sentence_1</code>

• Approximate statistics based on the first 36 samples:

  	sentence_0	sentence_1
  type	string	string
  details	<ul><li>min: 11 tokens</li><li>mean: 17.5 tokens</li><li>max: 24 tokens</li></ul>	<ul><li>min: 79 tokens</li><li>mean: 266.47 tokens</li><li>max: 331 tokens</li></ul>

• Samples:

  sentence_0	sentence_1
  <code>How may a complaint be made to the Commissioner under the Privacy Act?</code>	<code>Part V Investigations etc.
  Division 1 Investigation of complaints and investigations on the Commissioner’s
  initiative

  Section 36

  258 Privacy Act 1988
  Compilation No. 104 Compilation date: 04/06/2026

  Division 1—Investigation of complaints and investigations on the Commissioner’s initiative 36 Complaints (1) An individual may complain to the Commissioner about an act or practice that may be an interference with the privacy of the individual. (2) In the case of an act or practice that may be an interference with the privacy of 2 or more individuals, any one of those individuals may make a complaint under subsection (1) on behalf of all of the individuals. (2A) In the case of a representative complaint, this section has effect subject to section 38. (3) A complaint shall be in writing. (4) It is the duty of members of the staff of the Commissioner to provide appropriate assistance to a person who: (a) wishes to make a complaint; and (b) requires assi...</code> | | <code>What matters must an APP entity's notification under APP 5 ordinarily cover?</code> | <code>Australian Privacy Principles Schedule 1 Collection of personal information Part 2

  Clause 5

  Privacy Act 1988 367 Compilation No. 104 Compilation date: 04/06/2026

  5 Australian Privacy Principle 5—notification of the collection of personal information 5.1 At or before the time or, if that is not practicable, as soon as practicable after, an APP entity collects personal information about an individual, the entity must take such steps (if any) as are reasonable in the circumstances: (a) to notify the individual of such matters referred to in subclause 5.2 as are reasonable in the circumstances; or (b) to otherwise ensure that the individual is aware of any such matters. 5.2 The matters for the purposes of subclause 5.1 are as follows: (a) the identity and contact details of the APP entity; (b) if: (i) the APP entity collects the personal information from someone other than the individual; or (ii) the individual may not be aware that the APP entity...</code> | | <code>How may an individual access credit eligibility information held about them by a credit provider?</code> | <code>Credit reporting Part IIIA Credit providers Division 3

  Section 21T

  Privacy Act 1988 153 Compilation No. 104 Compilation date: 04/06/2026

  (b) from unauthorised access, modification or disclosure. (2) If: (a) a credit provider holds credit eligibility information about an individual; and (b) the provider no longer needs the information for any purpose for which the information may be used or disclosed by the provider under this Division; and (c) the provider is not required by or under an Australian law, or a court/tribunal order, to retain the information; the provider must take such steps as are reasonable in the circumstances to destroy the information or to ensure that the information is de-identified. Civil penalty: 1,000 penalty units. (3) If a credit provider is an APP entity, Australian Privacy Principle 11 does not apply to the provider in relation to credit eligibility information. Subdivision F—Access to, and correction of, information 21T...</code> |

• Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• num_train_epochs: 8
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

<details><summary>Click to expand</summary>

• do_predict: False
• prediction_loss_only: True
• per_device_train_batch_size: 8
• per_device_eval_batch_size: 8
• 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
• num_train_epochs: 8
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: None
• warmup_ratio: None
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• enable_jit_checkpoint: False
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• use_cpu: False
• seed: 42
• data_seed: None
• bf16: False
• fp16: False
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: -1
• ddp_backend: None
• debug: []
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: []
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• parallelism_config: None
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch_fused
• optim_args: None
• group_by_length: False
• length_column_name: length
• project: huggingface
• trackio_space_id: trackio
• 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
• push_to_hub: False
• resume_from_checkpoint: None
• hub_model_id: None
• hub_strategy: every_save
• hub_private_repo: None
• hub_always_push: False
• hub_revision: None
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_for_metrics: []
• eval_do_concat_batches: True
• auto_find_batch_size: False
• full_determinism: False
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• include_num_input_tokens_seen: no
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• eval_on_start: False
• use_liger_kernel: False
• liger_kernel_config: None
• eval_use_gather_object: False
• average_tokens_across_devices: True
• use_cache: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: round_robin
• router_mapping: {}
• learning_rate_mapping: {}

</details>

Training Time

• Training: 18.6 seconds

Framework Versions

• Python: 3.12.13
• Sentence Transformers: 5.4.1
• Transformers: 5.0.0
• PyTorch: 2.10.0+cu128
• Accelerate: 1.13.0
• Datasets: 5.0.0
• Tokenizers: 0.22.2

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{oord2019representationlearningcontrastivepredictive,
      title={Representation Learning with Contrastive Predictive Coding},
      author={Aaron van den Oord and Yazhe Li and Oriol Vinyals},
      year={2019},
      eprint={1807.03748},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/1807.03748},
}

<!-- ## Glossary *Clearly define terms in order to be accessible across audiences.* --> <!-- ## Model Card Authors *Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.* --> <!-- ## Model Card Contact *Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.* -->