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

This is a sentence-transformers model finetuned from BAAI/bge-small-en-v1.5. It maps inputs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, classification, clustering, and more.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: BAAI/bge-small-en-v1.5
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 384 dimensions
• Similarity Function: Cosine Similarity
• Supported Modality: Text

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': 384, 'pooling_mode': 'cls', 'include_prompt': True})
  (2): Normalize({'module_input_name': 'sentence_embedding', 'module_output_name': 'sentence_embedding'})
)

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 = [
    'Represent this sentence for searching relevant passages: Which years does the table provide information for accounts receivable for the company?',
    '||March 31,||\n||2019|2018|\n|Trade accounts receivable|$875.8|$557.8|\n|Other|6.8|8.1|\n|Total accounts receivable, gross|882.6|565.9|\n|Less allowance for doubtful accounts|2.0|2.2|\n|Total accounts receivable, net|$880.6|$563.7|\nNote 8. Other Financial Statement Details\nAccounts Receivable\nAccounts receivable consists of the following (in millions):',
    '|Year Ended May 31,|||||\n||||Percent Change||\n|(Dollars in millions)|2019|Actual|Constant|2018|\n|Interest expense|$2,082|3%|3%|$2,025|\nInterest Expense:\nInterest expense increased in fiscal 2019 compared to fiscal 2018 primarily due to higher average borrowings resulting from our issuance of $10.0 billion of senior notes in November 2017, which was partially offset by a reduction in interest expense resulting primarily from the maturities and repayments of $2.0 billion of senior notes during fiscal 2019 and $6.0 billion of senior notes during fiscal 2018.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.6351, 0.1055],
#         [0.6351, 1.0000, 0.1877],
#         [0.1055, 0.1877, 1.0000]])

Training Details

Training Dataset

Unnamed Dataset

• Size: 8,457 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 100 samples:

  	anchor	positive
  type	string	string
  modality	text	text
  details	min: 19 tokensmean: 29.68 tokensmax: 54 tokens	min: 195 tokensmean: 491.32 tokensmax: 512 tokens

• Samples:

  anchor	positive
  Represent this sentence for searching relevant passages: what is the net change in net revenue during 2015 for entergy corporation?	entergy corporation and subsidiaries management 2019s financial discussion and analysis a result of the entergy louisiana and entergy gulf states louisiana business combination , results of operations for 2015 also include two items that occurred in october 2015 : 1 ) a deferred tax asset and resulting net increase in tax basis of approximately $ 334 million and 2 ) a regulatory liability of $ 107 million ( $ 66 million net-of-tax ) as a result of customer credits to be realized by electric customers of entergy louisiana , consistent with the terms of the stipulated settlement in the business combination proceeding . see note 2 to the financial statements for further discussion of the business combination and customer credits . results of operations for 2015 also include the sale in december 2015 of the 583 mw rhode island state energy center for a realized gain of $ 154 million ( $ 100 million net-of-tax ) on the sale and the $ 77 million ( $ 47 million net-of-tax ) write-off and regu...
  Represent this sentence for searching relevant passages: what percentage of total facilities as measured in square feet are leased?	item 1b . unresolved staff comments not applicable . item 2 . properties as of december 26 , 2015 , our major facilities consisted of : ( square feet in millions ) united states countries total owned facilities1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30.7 17.2 47.9 leased facilities2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1 6.0 8.1 .|( square feet in millions )|unitedstates|othercountries|total||owned facilities1|30.7|17.2|47.9||leased facilities2|2.1|6.0|8.1||total facilities|32.8|23.2|56.0|1 leases on portions of the land used for these facilities expire on varying dates through 2062 . 2 leases expire on varying dates through 2030 and generally include renewals at our option . our principal executive offices are located in the u.s . and a majority of our wafer fabrication activities are also located in the u.s . we completed ...
  Represent this sentence for searching relevant passages: what is the percentage change in cash flow hedges in 2011 compare to the 2010?	undesignated hedges was $ 41.2 million and $ 42.1 million , respectively . the fair value of these hedging instruments in the company 2019s consolidated balance sheets as of october 29 , 2011 and october 30 , 2010 was immaterial . interest rate exposure management 2014 on june 30 , 2009 , the company entered into interest rate swap transactions related to its outstanding 5.0% ( 5.0 % ) senior unsecured notes where the company swapped the notional amount of its $ 375 million of fixed rate debt at 5.0% ( 5.0 % ) into floating interest rate debt through july 1 , 2014 . under the terms of the swaps , the company will ( i ) receive on the $ 375 million notional amount a 5.0% ( 5.0 % ) annual interest payment that is paid in two installments on the 1st of every january and july , commencing january 1 , 2010 through and ending on the maturity date ; and ( ii ) pay on the $ 375 million notional amount an annual three month libor plus 2.05% ( 2.05 % ) ( 2.42% ( 2.42 % ) as of october 29 , 2011 ...

• Loss: MultipleNegativesRankingLoss 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

• per_device_train_batch_size: 64
• num_train_epochs: 4
• learning_rate: 2e-05
• warmup_steps: 0.1
• fp16: True
• batch_sampler: no_duplicates

All Hyperparameters

• per_device_train_batch_size: 64
• num_train_epochs: 4
• max_steps: -1
• learning_rate: 2e-05
• lr_scheduler_type: linear
• lr_scheduler_kwargs: None
• warmup_steps: 0.1
• optim: adamw_torch_fused
• optim_args: None
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• optim_target_modules: None
• gradient_accumulation_steps: 1
• average_tokens_across_devices: True
• max_grad_norm: 1.0
• label_smoothing_factor: 0.0
• bf16: False
• fp16: True
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• use_liger_kernel: False
• liger_kernel_config: None
• use_cache: False
• neftune_noise_alpha: None
• torch_empty_cache_steps: None
• auto_find_batch_size: False
• log_on_each_node: True
• logging_nan_inf_filter: True
• include_num_input_tokens_seen: no
• log_level: passive
• log_level_replica: warning
• disable_tqdm: False
• project: huggingface
• trackio_space_id: None
• trackio_bucket_id: None
• trackio_static_space_id: None
• per_device_eval_batch_size: 8
• prediction_loss_only: True
• eval_on_start: False
• eval_do_concat_batches: True
• eval_use_gather_object: False
• eval_accumulation_steps: None
• include_for_metrics: []
• batch_eval_metrics: False
• save_only_model: False
• save_on_each_node: False
• enable_jit_checkpoint: False
• push_to_hub: False
• hub_private_repo: None
• hub_model_id: None
• hub_strategy: every_save
• hub_always_push: False
• hub_revision: None
• load_best_model_at_end: False
• ignore_data_skip: False
• restore_callback_states_from_checkpoint: False
• full_determinism: False
• seed: 42
• data_seed: None
• use_cpu: 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
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_pin_memory: True
• dataloader_persistent_workers: False
• dataloader_prefetch_factor: None
• dataloader_multiprocessing_context: None
• dataloader_in_order: True
• remove_unused_columns: True
• label_names: None
• train_sampling_strategy: random
• length_column_name: length
• ddp_find_unused_parameters: None
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• ddp_static_graph: None
• ddp_backend: None
• ddp_timeout: 1800
• fsdp: None
• fsdp_config: None
• deepspeed: None
• debug: []
• skip_memory_metrics: True
• do_predict: False
• resume_from_checkpoint: None
• local_rank: -1
• prompts: None
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}
• warmup_ratio: None

Training Logs

Training Time

• Training: 1.5 minutes

Framework Versions

• Python: 3.12.3
• Sentence Transformers: 6.0.1
• Transformers: 5.16.1
• PyTorch: 2.14.0+cu130
• Accelerate: 1.14.0
• Datasets: 5.0.1
• Tokenizers: 0.23.2

Additional Resources

• Training and Finetuning Embedding Models with Sentence Transformers: the end-to-end guide for training or finetuning Sentence Transformer models.
• Introduction to Matryoshka Embedding Models: variable-size embeddings that can be truncated with minimal quality loss.
• Binary and Scalar Embedding Quantization for Significantly Faster & Cheaper Retrieval: post-training compression of embedding vectors.
• Multimodal Embedding & Reranker Models with Sentence Transformers: use text, image, audio, and video models through the same API.
• Training and Finetuning Multimodal Embedding & Reranker Models with Sentence Transformers: train multimodal embedding models, with a Visual Document Retrieval walkthrough.

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},
}