pavanmantha/bge-base-en-bioembed768

BGE base BioASQ Matryoshka

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("pavanmantha/bge-base-en-bioembed768")
# Run inference
sentences = [
    "Basset is an open source package which applies CNNs to learn the functional activity of DNA sequences from genomics data. Basset was trained on a compendium of accessible genomic sites mapped in 164 cell types by DNase-seq, and demonstrated greater predictive accuracy than previous methods. Basset predictions for the change in accessibility between variant alleles were far greater for Genome-wide association study (GWAS) SNPs that are likely to be causal relative to nearby SNPs in linkage disequilibrium with them. With Basset, a researcher can perform a single sequencing assay in their cell type of interest and simultaneously learn that cell's chromatin accessibility code and annotate every mutation in the genome with its influence on present accessibility and latent potential for accessibility. Thus, Basset offers a powerful computational approach to annotate and interpret the noncoding genome.",
    'Describe the applicability of Basset in the context of deep learning',
    'What is the causative agent of the "Panama disease" affecting bananas?',
]
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	0.8432
cosine_accuracy@3	0.9428
cosine_accuracy@5	0.9619
cosine_accuracy@10	0.9788
cosine_precision@1	0.8432
cosine_precision@3	0.3143
cosine_precision@5	0.1924
cosine_precision@10	0.0979
cosine_recall@1	0.8432
cosine_recall@3	0.9428
cosine_recall@5	0.9619
cosine_recall@10	0.9788
cosine_ndcg@10	0.9168
cosine_mrr@10	0.8963
cosine_map@100	0.8972

Information Retrieval

• Dataset: dim_512
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.8538
cosine_accuracy@3	0.9428
cosine_accuracy@5	0.9619
cosine_accuracy@10	0.9746
cosine_precision@1	0.8538
cosine_precision@3	0.3143
cosine_precision@5	0.1924
cosine_precision@10	0.0975
cosine_recall@1	0.8538
cosine_recall@3	0.9428
cosine_recall@5	0.9619
cosine_recall@10	0.9746
cosine_ndcg@10	0.9198
cosine_mrr@10	0.9017
cosine_map@100	0.9027

Information Retrieval

• Dataset: dim_256
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.8453
cosine_accuracy@3	0.9386
cosine_accuracy@5	0.9555
cosine_accuracy@10	0.9746
cosine_precision@1	0.8453
cosine_precision@3	0.3129
cosine_precision@5	0.1911
cosine_precision@10	0.0975
cosine_recall@1	0.8453
cosine_recall@3	0.9386
cosine_recall@5	0.9555
cosine_recall@10	0.9746
cosine_ndcg@10	0.9142
cosine_mrr@10	0.8945
cosine_map@100	0.8953

Information Retrieval

• Dataset: dim_128
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.822
cosine_accuracy@3	0.928
cosine_accuracy@5	0.9449
cosine_accuracy@10	0.9703
cosine_precision@1	0.822
cosine_precision@3	0.3093
cosine_precision@5	0.189
cosine_precision@10	0.097
cosine_recall@1	0.822
cosine_recall@3	0.928
cosine_recall@5	0.9449
cosine_recall@10	0.9703
cosine_ndcg@10	0.9015
cosine_mrr@10	0.879
cosine_map@100	0.8801

Training Details

Training Dataset

Unnamed Dataset

• Size: 4,247 training samples
• Columns: positive and anchor
• Approximate statistics based on the first 1000 samples:

  	positive	anchor
  type	string	string
  details	min: 3 tokens mean: 102.44 tokens max: 512 tokens	min: 5 tokens mean: 15.78 tokens max: 44 tokens

• Samples:

  positive	anchor
  Restless legs syndrome (RLS), also known as Willis-Ekbom disease (WED), is a common movement disorder characterized by an uncontrollable urge to move because of uncomfortable, sometimes painful sensations in the legs with a diurnal variation and a release with movement.	Willis-Ekbom disease is also known as?
  Report the outcomes of laser in situ keratomileusis (LASIK) for high myopia correction after long-term follow-up['Report the outcomes of laser in situ keratomileusis (LASIK) for high myopia correction after long-term follow-up.']Laser in situ keratomileusis is also known as LASIKLaser in situ keratomileusis (LASIK)	What is another name for keratomileusis?
  CellMaps is an HTML5 open-source web tool that allows displaying, editing, exploring and analyzing biological networks as well as integrating metadata into them.CellMaps is an HTML5 open-source web tool that allows displaying, editing, exploring and analyzing biological networks as well as integrating metadata into them. CellMaps can easily be integrated in any web page by using an available JavaScript API. Computations and analyses are remotely executed in high-end servers, and all the functionalities are available through RESTful web services. CellMaps is an HTML5 open-source web tool that allows displaying, editing, exploring and analyzing biological networks as well as integrating metadata into them. Computations and analyses are remotely executed in high-end servers, and all the functionalities are available through RESTful web services. CellMaps can easily be integrated in any web page by using an available JavaScript API. CellMaps is an HTML5 open-source web tool that allows displaying, editing, exploring and analyzing biological networks as well as integrating metadata into them. Computations and analyses are remotely executed in high-end servers, and all the functionalities are available through RESTful web services. CellMaps can easily be integrated in any web page by using an available JavaScript API.CellMaps is an HTML5 open-source web tool that allows displaying, editing, exploring and analyzing biological networks as well as integrating metadata into them. CellMaps is an HTML5 open-source web tool that allows displaying, editing, exploring and analyzing biological networks as well as integrating metadata into them. CellMaps can easily be integrated in any web page by using an available JavaScript API. Computations and analyses are remotely executed in high-end servers, and all the functionalities are available through RESTful web services.	What is CellMaps?

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          128
      ],
      "matryoshka_weights": [
          1,
          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
• gradient_accumulation_steps: 16
• learning_rate: 2e-05
• num_train_epochs: 10
• lr_scheduler_type: cosine
• warmup_ratio: 0.1
• fp16: True
• tf32: False
• 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: 16
• 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: 10
• 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: False
• fp16: True
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: False
• 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_128_cosine_map@100	dim_256_cosine_map@100	dim_512_cosine_map@100	dim_768_cosine_map@100
0.9624	8	-	0.8560	0.8821	0.8904	0.8876
1.2030	10	1.2833	-	-	-	-
1.9248	16	-	0.8655	0.8808	0.8909	0.8889
2.4060	20	0.4785	-	-	-	-
2.8872	24	-	0.8720	0.8875	0.8893	0.8921
3.6090	30	0.2417	-	-	-	-
3.9699	33	-	0.8751	0.8924	0.8955	0.8960
4.8120	40	0.1607	-	-	-	-
4.9323	41	-	0.8799	0.8932	0.8964	0.8952
5.8947	49	-	0.8785	0.8944	0.9009	0.8982
6.0150	50	0.1152	-	-	-	-
6.9774	58	-	0.8803	0.8947	0.9018	0.8975
7.2180	60	0.0924	-	-	-	-
7.9398	66	-	0.8802	0.8956	0.9016	0.8973
8.4211	70	0.0832	-	-	-	-
8.9023	74	-	0.8801	0.8956	0.9027	0.8972
9.6241	80	0.074	0.8801	0.8953	0.9027	0.8972

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.10.13
• Sentence Transformers: 3.0.1
• Transformers: 4.41.2
• PyTorch: 2.1.2
• Accelerate: 0.31.0
• Datasets: 2.19.2
• 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}
}