SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2. 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: sentence-transformers/all-MiniLM-L6-v2
• Maximum Sequence Length: 256 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': 'mean', '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 = [
    'Mitochondria play a major role in apoptosis.',
    'Mitochondria: Dynamic Organelles in Disease, Aging, and Development\nMitochondria are the primary energy-generating system in most eukaryotic cells. Additionally, they participate in intermediary metabolism, calcium signaling, and apoptosis. Given these well-established functions, it might be expected that mitochondrial dysfunction would give rise to a simple and predictable set of defects in all tissues. However, mitochondrial dysfunction has pleiotropic effects in multicellular organisms. Clearly, much about the basic biology of mitochondria remains to be understood. Here we discuss recent work that suggests that the dynamics (fusion and fission) of these organelles is important in development and disease.',
    "The EMBL nucleotide sequence database.\nThe EMBL Nucleotide Sequence Database (http://www.ebi.ac.uk/embl. html ) constitutes Europe's primary nucleotide sequence resource. DNA and RNA sequences are directly submitted from researchers and genome sequencing groups and collected from the scientific literature and patent applications (Fig. 1). In collaboration with DDBJ and GenBank the database is produced, maintained and distributed at the European Bioinformatics Institute. Database releases are produced quarterly and are distributed on CD-ROM. EBI's network services allow access to the most up-to-date data collection via Internet and World Wide Web interface, providing database searching and sequence similarity facilities plus access to a large number of additional databases.",
]
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,  0.6941, -0.0084],
#         [ 0.6941,  1.0000,  0.0061],
#         [-0.0084,  0.0061,  1.0000]])

Training Details

Training Dataset

Unnamed Dataset

• Size: 733 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 100 samples:

  	anchor	positive
  type	string	string
  modality	text	text
  details	min: 11 tokensmean: 22.61 tokensmax: 43 tokens	min: 126 tokensmean: 246.84 tokensmax: 256 tokens

• Samples:

  anchor	positive
  0-dimensional biomaterials lack inductive properties.	New opportunities: the use of nanotechnologies to manipulate and track stem cells.Nanotechnologies are emerging platforms that could be useful in measuring, understanding, and manipulating stem cells. Examples include magnetic nanoparticles and quantum dots for stem cell labeling and in vivo tracking; nanoparticles, carbon nanotubes, and polyplexes for the intracellular delivery of genes/oligonucleotides and protein/peptides; and engineered nanometer-scale scaffolds for stem cell differentiation and transplantation. This review examines the use of nanotechnologies for stem cell tracking, differentiation, and transplantation. We further discuss their utility and the potential concerns regarding their cytotoxicity.
  R2D2 stops miRNA production by increasing the selectivity of Dcr2 for long dsRNA.	Phosphate and R2D2 restrict the substrate specificity of Dicer-2, an ATP-driven ribonuclease.Drosophila Dicer-2 generates small interfering RNAs (siRNAs) from long double-stranded RNA (dsRNA), whereas Dicer-1 produces microRNAs (miRNAs) from pre-miRNA. What makes the two Dicers specific for their biological substrates? We find that purified Dicer-2 can efficiently cleave pre-miRNA, but that inorganic phosphate and the Dicer-2 partner protein R2D2 inhibit pre-miRNA cleavage. Dicer-2 contains C-terminal RNase III domains that mediate RNA cleavage and an N-terminal helicase motif, whose function is unclear. We show that Dicer-2 is a dsRNA-stimulated ATPase that hydrolyzes ATP to ADP; ATP hydrolysis is required for Dicer-2 to process long dsRNA, but not pre-miRNA. Wild-type Dicer-2, but not a mutant defective in ATP hydrolysis, can generate siRNAs faster than it can dissociate from a long dsRNA substrate. We propose that the Dicer-2 helicase domain uses ATP to generate many siRNAs from a ...
  RA activation of DIF2 and NB4 cells induces hallmarks of transcriptionally active promoters.	Histone Methylation-Dependent Mechanisms Impose Ligand Dependency for Gene Activation by Nuclear ReceptorsNuclear receptors undergo ligand-dependent conformational changes that are required for corepressor-coactivator exchange, but whether there is an actual requirement for specific epigenetic landmarks to impose ligand dependency for gene activation remains unknown. Here we report an unexpected and general strategy that is based on the requirement for specific cohorts of inhibitory histone methyltransferases (HMTs) to impose gene-specific gatekeeper functions that prevent unliganded nuclear receptors and other classes of regulated transcription factors from binding to their target gene promoters and causing constitutive gene activation in the absence of stimulating signals. This strategy, based at least in part on an HMT-dependent inhibitory histone code, imposes a requirement for specific histone demethylases, including LSD1, to permit ligand- and signal-dependent activation of regu...

• 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: 16
• num_train_epochs: 1
• learning_rate: 2e-05
• warmup_steps: 0.1
• weight_decay: 0.01
• data_seed: 42
• dataloader_pin_memory: False

All Hyperparameters

• per_device_train_batch_size: 16
• num_train_epochs: 1
• 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: 0.01
• 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: False
• 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: 42
• 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: False
• 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
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}
• warmup_ratio: None

Training Logs

Training Time

• Training: 6.3 minutes

Framework Versions

• Python: 3.11.16
• Sentence Transformers: 6.0.0
• Transformers: 5.15.0
• PyTorch: 2.13.0+cpu
• Accelerate: 1.14.0
• Datasets: 5.0.1
• Tokenizers: 0.22.2

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},
}