philipk22/ind312-ft-v0

SentenceTransformer based on Snowflake/snowflake-arctic-embed-m

This is a sentence-transformers model finetuned from Snowflake/snowflake-arctic-embed-m. 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: Snowflake/snowflake-arctic-embed-m
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity

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': False}) 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("philipk22/ind312-ft-v0")
# Run inference
sentences = [
    'What regulatory framework does 21 CFR Part 312 pertain to as of January 23, 2025?',
    '§ 312.315 Intermediate-size patient populations.\n21 CFR Part 312 (up to date as of 1/23/2025)\nInvestigational New Drug Application 21 CFR Part 312 (Jan. 23, 2025)\n21 CFR Part 312 (Jan. 23, 2025) (enhanced display) page 2 of 54',
    'risk-benefit judgment in making the final decision on approvability. As part of this evaluation, consistent\nwith the statement of purpose in § 312.80, FDA will consider whether the benefits of the drug outweigh\nthe known and potential risks of the drug and the need to answer remaining questions about risks and\nbenefits of the drug, taking into consideration the severity of the disease and the absence of satisfactory\nalternative therapy.',
]
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

• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.92
cosine_accuracy@3	0.99
cosine_accuracy@5	0.99
cosine_accuracy@10	1.0
cosine_precision@1	0.92
cosine_precision@3	0.33
cosine_precision@5	0.198
cosine_precision@10	0.1
cosine_recall@1	0.92
cosine_recall@3	0.99
cosine_recall@5	0.99
cosine_recall@10	1.0
cosine_ndcg@10	0.9638
cosine_mrr@10	0.9517
cosine_map@100	0.9517

Training Details

Training Dataset

Unnamed Dataset

• Size: 798 training samples
• Columns: sentence_0 and sentence_1
• Approximate statistics based on the first 798 samples:

  	sentence_0	sentence_1
  type	string	string
  details	min: 12 tokens mean: 20.82 tokens max: 46 tokens	min: 19 tokens mean: 93.06 tokens max: 158 tokens

• Samples:

  sentence_0	sentence_1
  What is the scope of Part 312 in Title 21 regarding investigational new drug applications?	Title 21 —Food and Drugs Chapter I —Food and Drug Administration, Department of Health and Human Services Subchapter D —Drugs for Human Use Part 312 Investigational New Drug Application Subpart A General Provisions § 312.1 Scope. § 312.2 Applicability. § 312.3 Definitions and interpretations. § 312.6 Labeling of an investigational new drug. § 312.7 Promotion of investigational drugs. § 312.8 Charging for investigational drugs under an IND. § 312.10 Waivers.
  How does § 3126 address the labeling requirements for investigational new drugs?	Title 21 —Food and Drugs Chapter I —Food and Drug Administration, Department of Health and Human Services Subchapter D —Drugs for Human Use Part 312 Investigational New Drug Application Subpart A General Provisions § 312.1 Scope. § 312.2 Applicability. § 312.3 Definitions and interpretations. § 312.6 Labeling of an investigational new drug. § 312.7 Promotion of investigational drugs. § 312.8 Charging for investigational drugs under an IND. § 312.10 Waivers.
  What are the general principles outlined in § 31222 regarding the IND submission?	§ 312.10 Waivers. Subpart B Investigational New Drug Application (IND) § 312.20 Requirement for an IND. § 312.21 Phases of an investigation. § 312.22 General principles of the IND submission. § 312.23 IND content and format. § 312.30 Protocol amendments. § 312.31 Information amendments. § 312.32 IND safety reporting. § 312.33 Annual reports. § 312.38 Withdrawal of an IND. Subpart C Administrative Actions

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          128,
          64
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 10
• per_device_eval_batch_size: 10
• num_train_epochs: 10
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 10
• per_device_eval_batch_size: 10
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• 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: 10
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.0
• 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: False
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• 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: False
• 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
• 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: None
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• include_for_metrics: []
• 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
• eval_on_start: False
• use_liger_kernel: False
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: round_robin

Training Logs

Epoch	Step	Training Loss	cosine_ndcg@10
0.625	50	-	0.9091
1.0	80	-	0.9209
1.25	100	-	0.9329
1.875	150	-	0.9439
2.0	160	-	0.9379
2.5	200	-	0.9367
3.0	240	-	0.9459
3.125	250	-	0.9432
3.75	300	-	0.9479
4.0	320	-	0.9515
4.375	350	-	0.9509
5.0	400	-	0.9581
5.625	450	-	0.9551
6.0	480	-	0.9604
6.25	500	0.3078	0.9577
6.875	550	-	0.9651
7.0	560	-	0.9651
7.5	600	-	0.9641
8.0	640	-	0.9641
8.125	650	-	0.9638
8.75	700	-	0.9638
9.0	720	-	0.9638
9.375	750	-	0.9601
10.0	800	-	0.9638

Framework Versions

• Python: 3.11.11
• Sentence Transformers: 3.4.1
• Transformers: 4.48.3
• PyTorch: 2.5.1+cu124
• Accelerate: 1.3.0
• Datasets: 3.3.2
• Tokenizers: 0.21.0

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}
}