bdanko/umsb-mpnet-episodic-memory

SentenceTransformer based on sentence-transformers/all-mpnet-base-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-mpnet-base-v2 on the umsb-episodic-memory dataset. 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: sentence-transformers/all-mpnet-base-v2
• Maximum Sequence Length: 384 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Supported Modality: Text
• Training Dataset:
  • umsb-episodic-memory

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': 'MPNetModel'})
  (1): Pooling({'embedding_dimension': 768, 'pooling_mode': 'mean', '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("sentence_transformers_model_id")
# Run inference
sentences = [
    'e medication administration date. This parameter corresponds to the dosageInstruction.timing.repeat.boundsPeriod element\nMedication orders that do not have start and end dates within the search parameter dates are filtered. If the environment supports multiple time zones, the search dates are adjusted one day in both directions, so more medications might be returned than expected. Use caution when filtering a medication list by date as it is possible to filter out important active medications. Starting in the November 2022 version of Epic, this parameter is respected',
    'e medication administration date. This parameter corresponds to the dosageInstruction.timing.repeat.boundsPeriod element\nMedication orders that do not have start and end dates within the search parameter dates are filtered. If the environment supports multiple time zones, the search dates are adjusted one day in both directions, so more medications might be returned than expected. Use caution when filtering a medication list by date as it is possible to filter out important active medications. Starting in the November 2022 version of Epic, this parameter is respected',
    'e": {"properties": {"text": {"description": "The flowsheet ID, encoded flowsheet ID, or LOINC codes to flowsheet mapping\nWhat is being measured.", "type": "string"}}, "type": "object"}, "effectiveDateTime": {"description": "The date and time the observation was taken, in ISO format.", "type": "string"}, "resourceType": {"description": "Use \\"Observation\\" for vitals observations.", "type": "string"}, "status": {"description": "The status of the observation. Only a value of \\"final\\" is supported',
]
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, 1.0000, 0.3683],
#         [1.0000, 1.0000, 0.3683],
#         [0.3683, 0.3683, 1.0000]])

Training Details

Training Dataset

umsb-episodic-memory

• Dataset: umsb-episodic-memory
• Size: 3,067 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 33 tokens mean: 221.24 tokens max: 384 tokens	min: 33 tokens mean: 221.24 tokens max: 384 tokens

• Samples:

  anchor	positive
  description": "The patient's family (last) name.", "type": "string"}, "gender": {"description": "The patient's legal sex Starting in the August 2021 version of Epic, the legal-sex parameter is preferred.", "type": "string"}, "given": {"description": "The patient's given name. May include first and middle names.", "type": "string"}, "identifier": {"description": "The patient's identifier.", "type": "string"}, "legal-sex": {"description": "The patient\u2019s legal sex. Takes precedence over the gender search parameter. Available starting in the August 2021 version of Epic.", "type": "string"}, "name": {"description": "Any part of the patient's name. When discrete name parameters are used, such as family or given, this parameter is ignored.", "type": "string"}, "telecom": {"description": "The patient's phone number or email.", "type": "string"}}, "required": [], "type": "object"}}]	description": "The patient's family (last) name.", "type": "string"}, "gender": {"description": "The patient's legal sex Starting in the August 2021 version of Epic, the legal-sex parameter is preferred.", "type": "string"}, "given": {"description": "The patient's given name. May include first and middle names.", "type": "string"}, "identifier": {"description": "The patient's identifier.", "type": "string"}, "legal-sex": {"description": "The patient\u2019s legal sex. Takes precedence over the gender search parameter. Available starting in the August 2021 version of Epic.", "type": "string"}, "name": {"description": "Any part of the patient's name. When discrete name parameters are used, such as family or given, this parameter is ignored.", "type": "string"}, "telecom": {"description": "The patient's phone number or email.", "type": "string"}}, "required": [], "type": "object"}}]
  e to filter out important active medications. Starting in the November 2022 version of Epic, this parameter is respected
  In May 2022 and earlier versions of Epic, this parameter is allowed but is ignored and no date filtering is applied.", "type": "string"}, "patient": {"description": "The FHIR patient ID.", "type": "string"}}, "required": ["patient"], "type": "object"}}, {"description": "MedicationRequest.Create", "name": "POST {api_base}/MedicationRequest", "parameters": {"properties": {"authoredOn": {"description": "The date the prescription was written.", "type": "string"}, "dosageInstruction": {"items": {"properties": {"doseAndRate": {"items": {"properties": {"doseQuantity": {"properties": {"unit": {"description": "unit for the dose such as "g" ", "type": "string"}, "value": {"type": "number"}}, "type": "object"}, "rateQuantity": {"properties": {"unit": {"description": "unit for the rate such as "h" ", "type": "string"}, "value": {"type": "number"}}, "type": "object"}}, "type"...	e to filter out important active medications. Starting in the November 2022 version of Epic, this parameter is respected
  In May 2022 and earlier versions of Epic, this parameter is allowed but is ignored and no date filtering is applied.", "type": "string"}, "patient": {"description": "The FHIR patient ID.", "type": "string"}}, "required": ["patient"], "type": "object"}}, {"description": "MedicationRequest.Create", "name": "POST {api_base}/MedicationRequest", "parameters": {"properties": {"authoredOn": {"description": "The date the prescription was written.", "type": "string"}, "dosageInstruction": {"items": {"properties": {"doseAndRate": {"items": {"properties": {"doseQuantity": {"properties": {"unit": {"description": "unit for the dose such as "g" ", "type": "string"}, "value": {"type": "number"}}, "type": "object"}, "rateQuantity": {"properties": {"unit": {"description": "unit for the rate such as "h" ", "type": "string"}, "value": {"type": "number"}}, "type": "object"}}, "type"...
  e": {"properties": {"text": {"description": "The flowsheet ID, encoded flowsheet ID, or LOINC codes to flowsheet mapping What is being measured.", "type": "string"}}, "type": "object"}, "effectiveDateTime": {"description": "The date and time the observation was taken, in ISO format.", "type": "string"}, "resourceType": {"description": "Use "Observation" for vitals observations.", "type": "string"}, "status": {"description": "The status of the observation. Only a value of "final" is supported	e": {"properties": {"text": {"description": "The flowsheet ID, encoded flowsheet ID, or LOINC codes to flowsheet mapping What is being measured.", "type": "string"}}, "type": "object"}, "effectiveDateTime": {"description": "The date and time the observation was taken, in ISO format.", "type": "string"}, "resourceType": {"description": "Use "Observation" for vitals observations.", "type": "string"}, "status": {"description": "The status of the observation. Only a value of "final" is supported

• 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: 32
• num_train_epochs: 2
• learning_rate: 2e-05
• warmup_steps: 0.05
• fp16: True

All Hyperparameters

Click to expand

• per_device_train_batch_size: 32
• num_train_epochs: 2
• max_steps: -1
• learning_rate: 2e-05
• lr_scheduler_type: linear
• lr_scheduler_kwargs: None
• warmup_steps: 0.05
• 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
• 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: []
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• deepspeed: None
• debug: []
• skip_memory_metrics: True
• do_predict: False
• resume_from_checkpoint: None
• warmup_ratio: None
• local_rank: -1
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss
0.2604	25	0.3777
0.5208	50	0.3348
0.7812	75	0.3936
1.0417	100	0.3583
1.3021	125	0.3227
1.5625	150	0.3224
1.8229	175	0.3492

Training Time

• Training: 2.4 minutes

Framework Versions

• Python: 3.11.12
• Sentence Transformers: 5.4.1
• Transformers: 5.8.0
• PyTorch: 2.11.0+cu130
• Accelerate: 1.13.0
• Datasets: 4.8.5
• 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},
}