Add new SentenceTransformer model

17ff38c verified 10 months ago

28.7 kB

metadata

language:
  - en
license: apache-2.0
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:5822
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
base_model: nomic-ai/modernbert-embed-base
widget:
  - source_sentence: >-
      VCH qualifies as a service-disabled veteran-owned small business and is an
      actual offeror on the 

      SDVOSB Pool Solicitation under the Polaris Program.  VCH Compl. ¶¶ 2, 5. 
      Both SHS and VCH 

      also claim to be prospective offerors on the SB Pool Solicitation.  SHS
      Compl. ¶ 5; VCH Compl. 

      ¶ 5.  Both SHS and VCH state they have prepared, but not yet formally
      submitted, proposals in
    sentences:
      - >-
        What type of request must be submitted according to the information
        security procedures?
      - On which solicitation is VCH an actual offeror?
      - >-
        According to which U.S. Code section is the term 'infrastructure
        security information' defined?
  - source_sentence: >-
      Count Nine in No. 11-444: May 13, 2010 FOIA Request to the CIA 

      First, the CIA refused to process the plaintiff’s FOIA request which
      sought “a 

      representative sample of [CIA] analytical reports and memoranda presenting
      psychological 

      analyses or profiles of foreign government officials, terrorist leaders,
      international criminals,
    sentences:
      - Under what section were 'Regional Boards' considered 'agencies'?
      - >-
        What is the count number associated with the May 13, 2010 FOIA Request
        to the CIA?
      - What type of contracts does FAR 37.102(a) most prefer?
  - source_sentence: >-
      records referencing FOIA and Privacy Act requests submitted by [ten listed
      parties] that contain 

      remarks, comments, notes, explanations, etc. made by CIA personnel or
      contractors about the 

      processing of these requests (and appeals, if appropriate), the
      invocations of exemptions, or 

      related matters.”  See Decl. of Martha M. Lutz (Sept. 26, 2012) (“Third
      Lutz Decl.”) Ex. A at 1,
    sentences:
      - What is mentioned as possibly contradicting the defendant's statement?
      - On what date is the document 'Third Lutz Decl.' declared?
      - What does the plaintiff argue regarding the phrase 'on notice'?
  - source_sentence: >-
      “[a]gencies shall evaluate all proposals in accordance with 15.305(a),
      and, if discussions are to be 

      conducted, establish the competitive range.  Based on the ratings of each
      proposal against all 

      evaluation criteria, the contracting officer shall establish a competitive
      range comprised of all of 

      the most highly rated proposals . . . .”  FAR 15.306(c)(1) (emphasis
      added).  This last provision
    sentences:
      - Who establishes the competitive range based on proposal ratings?
      - What type of basis are the services to be acquired on?
      - >-
        What is the source of the statement quoted about the agencies and
        advisory committees?
  - source_sentence: >-
      for a specific procurement through separate joint ventures with different
      protégés.”  Id.  The SBA 

      underscored this purpose by highlighting that in acquiring a second
      protégé, the mentor “has 

      already assured SBA that the two protégés would not be competitors.  If
      the two mentor-protégé 

      relationships were approved in the same [North American Industry
      Classification System] code,
    sentences:
      - >-
        Where can the details of the CIA's framing of the plaintiff's injury be
        found?
      - What is the context of the mentor-protégé relationships mentioned?
      - >-
        What requirement does the FOIA have regarding the release of segregable
        portions of a record?
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - cosine_accuracy@1
  - cosine_accuracy@3
  - cosine_accuracy@5
  - cosine_accuracy@10
  - cosine_precision@1
  - cosine_precision@3
  - cosine_precision@5
  - cosine_precision@10
  - cosine_recall@1
  - cosine_recall@3
  - cosine_recall@5
  - cosine_recall@10
  - cosine_ndcg@10
  - cosine_mrr@10
  - cosine_map@100
model-index:
  - name: ModernBERT Embed base Legal Matryoshka
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 768
          type: dim_768
        metrics:
          - type: cosine_accuracy@1
            value: 0.5564142194744977
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.6058732612055642
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.6955177743431221
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7758887171561051
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.5564142194744977
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.5265327150953116
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.40216383307573417
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.23879443585780524
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.19770736733642452
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.5216383307573416
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.6432251416795467
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.755280783101494
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6620393079072384
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6046944383111306
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6446423018095604
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 512
          type: dim_512
        metrics:
          - type: cosine_accuracy@1
            value: 0.5440494590417311
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.5873261205564142
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.6846986089644513
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7604327666151468
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.5440494590417311
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.5105615662029882
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.3935085007727976
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.23431221020092735
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.1947449768160742
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.5066975785677487
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.6317619783616693
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7434312210200927
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6493066767347394
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.591142268344741
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6320199198507571
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 256
          type: dim_256
        metrics:
          - type: cosine_accuracy@1
            value: 0.5131375579598145
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.5548686244204019
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.6383307573415765
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7078825347758887
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.5131375579598145
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.4848016486347243
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.3712519319938176
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.2185471406491499
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.18019062339000513
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.47797527047913446
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.5926069036579084
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.693070582174137
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6070012758911372
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.5553402762444491
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.5963300240538415
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 128
          type: dim_128
        metrics:
          - type: cosine_accuracy@1
            value: 0.45904173106646057
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.5115919629057187
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.5873261205564142
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.6553323029366306
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.45904173106646057
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.43379701184956204
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.3372488408037094
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.2017001545595054
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.16679546625450797
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.4330242143225141
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.5422462648119527
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.6388459556929419
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.5563637742997848
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.50374622801207
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.546814426794716
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 64
          type: dim_64
        metrics:
          - type: cosine_accuracy@1
            value: 0.36476043276661513
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.40494590417310666
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.47449768160741884
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.5409582689335394
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.36476043276661513
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.3482740855229263
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.2741885625965997
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.16599690880989182
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.1281555899021123
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.34119010819165374
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.4383049974240082
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.5255023183925811
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.44955240085023923
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.40338436250337323
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.4452507749409145
            name: Cosine Map@100

ModernBERT Embed base Legal Matryoshka

This is a sentence-transformers model finetuned from nomic-ai/modernbert-embed-base on the json dataset. 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: nomic-ai/modernbert-embed-base
Maximum Sequence Length: 8192 tokens
Output Dimensionality: 768 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- json
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': 8192, 'do_lower_case': False}) with Transformer model: ModernBertModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, '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("chrisekwugum/modernbert-embed-base-legal-matryoshka-2")
# Run inference
sentences = [
    'for a specific procurement through separate joint ventures with different protégés.”  Id.  The SBA \nunderscored this purpose by highlighting that in acquiring a second protégé, the mentor “has \nalready assured SBA that the two protégés would not be competitors.  If the two mentor-protégé \nrelationships were approved in the same [North American Industry Classification System] code,',
    'What is the context of the mentor-protégé relationships mentioned?',
    "Where can the details of the CIA's framing of the plaintiff's injury be found?",
]
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

Datasets: dim_768, dim_512, dim_256, dim_128 and dim_64
Evaluated with InformationRetrievalEvaluator

Metric	dim_768	dim_512	dim_256	dim_128	dim_64
cosine_accuracy@1	0.5564	0.544	0.5131	0.459	0.3648
cosine_accuracy@3	0.6059	0.5873	0.5549	0.5116	0.4049
cosine_accuracy@5	0.6955	0.6847	0.6383	0.5873	0.4745
cosine_accuracy@10	0.7759	0.7604	0.7079	0.6553	0.541
cosine_precision@1	0.5564	0.544	0.5131	0.459	0.3648
cosine_precision@3	0.5265	0.5106	0.4848	0.4338	0.3483
cosine_precision@5	0.4022	0.3935	0.3713	0.3372	0.2742
cosine_precision@10	0.2388	0.2343	0.2185	0.2017	0.166
cosine_recall@1	0.1977	0.1947	0.1802	0.1668	0.1282
cosine_recall@3	0.5216	0.5067	0.478	0.433	0.3412
cosine_recall@5	0.6432	0.6318	0.5926	0.5422	0.4383
cosine_recall@10	0.7553	0.7434	0.6931	0.6388	0.5255
cosine_ndcg@10	0.662	0.6493	0.607	0.5564	0.4496
cosine_mrr@10	0.6047	0.5911	0.5553	0.5037	0.4034
cosine_map@100	0.6446	0.632	0.5963	0.5468	0.4453

Training Details

Training Dataset

json

Dataset: json
Size: 5,822 training samples
Columns: positive and anchor
Approximate statistics based on the first 1000 samples:
positive anchor
type string string
details
min: 28 tokens
mean: 97.21 tokens
max: 170 tokens

min: 7 tokens
mean: 16.7 tokens
max: 39 tokens

	positive	anchor
type	string	string
details	min: 28 tokens mean: 97.21 tokens max: 170 tokens	min: 7 tokens mean: 16.7 tokens max: 39 tokens

Samples:

positive	anchor
`Counts Seven, Nine, and Ten in No. 11-445: February 6, 2010 FOIA Requests to the CIA, State Department, and NSA On February 6, 2010, the plaintiff submitted three substantially identical FOIA requests—one to the CIA, one to the State Department, and one to the National Security Agency (“NSA”). The request to the CIA sought “all current training handbooks, manuals, guidelines,`	`What is the number associated with the case involving Counts Seven, Nine, and Ten?`
`The Government’s notion of a categorical principle stems mainly from a series of decisions in this District. Defs.’ Mem. at 14; Defs.’ Reply at 9 n.2. The first was Gates v. Schlesinger, 366 F. Supp. 797 (D.D.C. 1973), which stated that “an advisory committee is not an ‘agency.’” Id. at 799. Gates’s first rationale for this conclusion was that FACA “utilizes the definition of`	`From where does the Government's notion of a categorical principle mainly stem?`
`sort its incoming FOIA requests based on fee categories.” First Lutz Decl. ¶ 11. The CIA’s declarant also states that “this information [i.e., fee category] is not included in the electronic system,” though the CIA’s declarant also avers that “[f]ee category is not a mandatory field,” and thus “this information is often not included in a FOIA request record.” Id. The plaintiff focuses`	`According to the CIA's declarant, is fee category a mandatory field?`

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: epoch
per_device_train_batch_size: 32
per_device_eval_batch_size: 16
gradient_accumulation_steps: 16
learning_rate: 2e-05
num_train_epochs: 4
lr_scheduler_type: cosine
warmup_ratio: 0.1
bf16: 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
torch_empty_cache_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: 4
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: True
fp16: False
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: 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: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	dim_768_cosine_ndcg@10	dim_512_cosine_ndcg@10	dim_256_cosine_ndcg@10	dim_128_cosine_ndcg@10	dim_64_cosine_ndcg@10
0.8791	10	91.392	-	-	-	-	-
1.0	12	-	0.6238	0.6027	0.5669	0.5230	0.4009
1.7033	20	38.8819	-	-	-	-	-
2.0	24	-	0.6596	0.6423	0.5986	0.5491	0.4384
2.5275	30	28.6263	-	-	-	-	-
3.0	36	-	0.6615	0.6502	0.6058	0.5575	0.4486
3.3516	40	25.2135	-	-	-	-	-
3.7033	44	-	0.6620	0.6493	0.6070	0.5564	0.4496

The bold row denotes the saved checkpoint.

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: 3.3.1
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}
}