ngiometti/legal-ft-3

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

This is a sentence-transformers model finetuned from Snowflake/snowflake-arctic-embed-l. It maps sentences & paragraphs to a 1024-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-l
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 1024 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': 1024, '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("ngiometti/legal-ft-3")
# Run inference
sentences = [
    'What is the primary focus of Fluence as a continuing education organization?',
    'Name":"Fluence","Specter - Company ID":"621f973f972ef7e5d69c8085","Domain":"fluencetraining.com","LinkedIn - URL":"linkedin.com/company/fluencetraining","Description":"Fluence is a leading continuing education organization in psychedelic therapy.","Company Size":"11-50","Industries":["Education","HR and Recruiting","Health","Software"],"Title":"Advisor","Departments":["Other"],"Start Date":"2023-07-01","End Date":null,"Location":"New York City, New York, United States, United States","Is Current":true,"Job Order":17},{"Company Name":"VentureKit","Specter - Company ID":null,"Domain":"venturekit.com","LinkedIn - URL":"linkedin.com/company/venturekit","Description":"VentureKit publishes free guides to help entrepreneurs get things',
    'Education Level: Bachelor\'s Degree\nCurrent Position Title: CTO, Head of Research\nCurrent Position Company Name: Mursion\nCurrent Position Company Website: mursion.com\nPast Position Title: CEO and Co-founder\nPast Position Company Name: DNABLOCK\nPast Position Company Website: dnablock.com\nCurrent Tenure: 85.0\nAverage Tenure: 34.0\nLanguages: [{"Name":"Spanish","Proficiency Level":"Limited Working Proficiency"},{"Name":"Arabic","Proficiency Level":"Limited Working Proficiency"}]\nLinkedIn - Followers: 5022.0\nLinkedIn - Connections: 2997.0',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

# 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	0.7917
cosine_accuracy@3	0.9667
cosine_accuracy@5	0.975
cosine_accuracy@10	0.9833
cosine_precision@1	0.7917
cosine_precision@3	0.3222
cosine_precision@5	0.195
cosine_precision@10	0.0983
cosine_recall@1	0.7917
cosine_recall@3	0.9667
cosine_recall@5	0.975
cosine_recall@10	0.9833
cosine_ndcg@10	0.9019
cosine_mrr@10	0.8741
cosine_map@100	0.8749

Training Details

Training Dataset

Unnamed Dataset

• Size: 786 training samples
• Columns: sentence_0 and sentence_1
• Approximate statistics based on the first 786 samples:

  	sentence_0	sentence_1
  type	string	string
  details	min: 9 tokens mean: 17.2 tokens max: 33 tokens	min: 11 tokens mean: 218.92 tokens max: 464 tokens

• Samples:

  sentence_0	sentence_1
  What types of products has the individual built experience in, according to the context?	experience in building world class hardware and software products for consumer electronics, aerospace and enterprise software solutions. Proven track record of building big-data cloud computing software and analytic software platform with AI, Computer Vision and Machine Learning. Progressive, innovative and highly valued for aligning corporate strategies with market opportunities, translating goals into actionable plans, and providing leadership to multi-discipline, cross cultural teams.
  How does the individual align corporate strategies with market opportunities?	experience in building world class hardware and software products for consumer electronics, aerospace and enterprise software solutions. Proven track record of building big-data cloud computing software and analytic software platform with AI, Computer Vision and Machine Learning. Progressive, innovative and highly valued for aligning corporate strategies with market opportunities, translating goals into actionable plans, and providing leadership to multi-discipline, cross cultural teams.
  What is the company size of Diamanti?	- Company ID":"5e3a8f19040ca7b0c6f031bf","Domain":"diamanti.com","LinkedIn - URL":"linkedin.com/company/diamanti","Description":null,"Company Size":"51-200","Industries":["Consumer Products","Hardware","Information Technology","Internet Services","Software"],"Title":"Chief Operating Officer","Departments":["Senior Leadership","Operations"],"Start Date":"2018-11-01","End Date":"2021-07-01","Location":"San Jose, California, United States, United States","Is Current":false,"Job Order":4},{"Company Name":"Planet","Specter - Company ID":"5e3bc13c00c8f4c966a7da4c","Domain":"planet.com","LinkedIn - URL":"linkedin.com/company/planet-labs","Description":"Planet operates the world's largest fleet of Earth imaging satellites to daily image the entire

• 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
0.6329	50	-	0.8917
1.0	79	-	0.9080
1.2658	100	-	0.9265
1.8987	150	-	0.9091
2.0	158	-	0.9100
2.5316	200	-	0.9214
3.0	237	-	0.9110
3.1646	250	-	0.9161
3.7975	300	-	0.9108
4.0	316	-	0.9145
4.4304	350	-	0.8955
5.0	395	-	0.9019
5.0633	400	-	0.9008
5.6962	450	-	0.8980
6.0	474	-	0.9036
6.3291	500	0.7603	0.9021
6.9620	550	-	0.8977
7.0	553	-	0.8976
7.5949	600	-	0.9059
8.0	632	-	0.9005
8.2278	650	-	0.9039
8.8608	700	-	0.9050
9.0	711	-	0.9052
9.4937	750	-	0.9021
10.0	790	-	0.9019

Framework Versions

• Python: 3.13.1
• Sentence Transformers: 3.4.1
• Transformers: 4.49.0
• PyTorch: 2.6.0+cu124
• Accelerate: 1.4.0
• 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}
}