metadata

language:
  - en
license: apache-2.0
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:6300
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
base_model: BAAI/bge-base-en-v1.5
widget:
  - source_sentence: >-
      The fair value of consideration transferred of $212.1 million consisted
      of: (1) cash consideration paid of $211.3 million, net of cash acquired,
      and (2) non-cash consideration of $0.8 million representing the portion of
      the replacement equity awards issued in connection with the acquisition
      that was associated with services rendered through the date of the
      acquisition.
    sentences:
      - >-
        What is the monthly cost of a Connected Fitness Subscription if it
        includes a combination of a Bike, Tread, Guide, or Row product in the
        same household as of June 2022?
      - >-
        What was the fair value of the total consideration transferred for the
        acquisition discussed, and how was it composed?
      - >-
        How did the Tax Court rule on November 18, 2020, regarding the company's
        dispute with the IRS?
  - source_sentence: >-
      Each of the UK LSA members has agreed, on a several and not joint basis,
      to compensate the Company for certain losses which may be incurred by the
      Company, Visa Europe or their affiliates as a result of certain existing
      and potential litigation relating to the setting and implementation of
      domestic multilateral interchange fee rates in the United Kingdom prior to
      the closing of the Visa Europe acquisition (Closing), subject to the terms
      and conditions set forth therein and, with respect to each UK LSA member,
      up to a maximum amount of the up-front cash consideration received by such
      UK LSA member. The UK LSA members’ obligations under the UK loss sharing
      agreement are conditional upon, among other things, either (a) losses
      valued in excess of the sterling equivalent on June 21, 2016 of €1.0
      billion having arisen in UK covered claims (and such losses having reduced
      the conversion rate of the series B preferred stock accordingly), or (b)
      the conversion rate of the series B preferred stock having been reduced to
      zero pursuant to losses arising in claims...
    sentences:
      - >-
        Are AbbVie's corporate governance materials available to the public, and
        if so, where?
      - >-
        What conditions must be met for the UK loss sharing agreement to
        compensate for losses?
      - >-
        How much did Delta Air Lines recognize in government grants from the
        Payroll Support Programs during the year ended December 31, 2021?
  - source_sentence: >-
      We provide our customers with an opportunity to trade-in their pre-owned
      gaming, mobility, and other products at our stores in exchange for cash or
      credit which can be applied towards the purchase of other products.
    sentences:
      - What is GameStop's trade-in program?
      - >-
        What were the total unrealized losses on U.S. Treasury securities as of
        the last reporting date?
      - >-
        What methods can a refinery use to meet its Environmental Protection
        Agency (EPA) requirements for blending renewable fuels?
  - source_sentence: >-
      Diluted earnings per share is calculated using our weighted-average
      outstanding common shares including the dilutive effect of stock awards as
      determined under the treasury stock method.
    sentences:
      - >-
        How do changes in the assumed long-term rate of return affect AbbVie's
        net periodic benefit cost for pension plans?
      - >-
        What are the primary factors discussed in the Management’s Discussion
        and Analysis that affect the financial statements year-to-year changes?
      - What is the method used to calculate diluted earnings per share?
  - source_sentence: >-
      Item 8 in the document covers 'Financial Statements and Supplementary
      Data'.
    sentences:
      - What type of information does Item 8 in the document cover?
      - >-
        What are some of the potential consequences for Meta Platforms, Inc.
        from inquiries or investigations as noted in the provided text?
      - How is the take rate calculated and what does it represent?
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: BGE base Financial Matryoshka
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 768
          type: dim_768
        metrics:
          - type: cosine_accuracy@1
            value: 0.68
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.8242857142857143
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.8571428571428571
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8985714285714286
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.68
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.27476190476190476
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.1714285714285714
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.08985714285714284
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.68
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.8242857142857143
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.8571428571428571
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8985714285714286
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.7931022011968226
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.759021541950113
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.7627727073081649
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 512
          type: dim_512
        metrics:
          - type: cosine_accuracy@1
            value: 0.6685714285714286
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.82
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.86
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.9042857142857142
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.6685714285714286
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.2733333333333333
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.172
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.09042857142857141
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.6685714285714286
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.82
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.86
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.9042857142857142
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.7907009828560375
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.7540430839002267
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.7572918009226873
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 256
          type: dim_256
        metrics:
          - type: cosine_accuracy@1
            value: 0.6771428571428572
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.8142857142857143
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.8571428571428571
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8857142857142857
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.6771428571428572
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.2714285714285714
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.1714285714285714
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.08857142857142855
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.6771428571428572
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.8142857142857143
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.8571428571428571
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8857142857142857
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.7870155634206691
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.7548027210884352
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.7592885578023618
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 128
          type: dim_128
        metrics:
          - type: cosine_accuracy@1
            value: 0.6542857142857142
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.8071428571428572
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.8514285714285714
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8857142857142857
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.6542857142857142
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.26904761904761904
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.17028571428571426
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.08857142857142856
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.6542857142857142
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.8071428571428572
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.8514285714285714
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8857142857142857
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.7751084647376248
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.73912925170068
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.7430473786684797
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 64
          type: dim_64
        metrics:
          - type: cosine_accuracy@1
            value: 0.6157142857142858
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7771428571428571
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.8214285714285714
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8728571428571429
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.6157142857142858
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.259047619047619
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.16428571428571428
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.08728571428571427
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.6157142857142858
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7771428571428571
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.8214285714285714
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8728571428571429
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.7472883962433147
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.7067517006802716
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.7111439006196084
            name: Cosine Map@100

BGE base Financial Matryoshka

This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5 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: BAAI/bge-base-en-v1.5
Maximum Sequence Length: 512 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': 512, 'do_lower_case': True}) 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("Shivam1311/bge-base-financial-matryoshka")
# Run inference
sentences = [
    "Item 8 in the document covers 'Financial Statements and Supplementary Data'.",
    'What type of information does Item 8 in the document cover?',
    'What are some of the potential consequences for Meta Platforms, Inc. from inquiries or investigations as noted in the provided text?',
]
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	0.68	0.6686	0.6771	0.6543	0.6157
cosine_accuracy@3	0.8243	0.82	0.8143	0.8071	0.7771
cosine_accuracy@5	0.8571	0.86	0.8571	0.8514	0.8214
cosine_accuracy@10	0.8986	0.9043	0.8857	0.8857	0.8729
cosine_precision@1	0.68	0.6686	0.6771	0.6543	0.6157
cosine_precision@3	0.2748	0.2733	0.2714	0.269	0.259
cosine_precision@5	0.1714	0.172	0.1714	0.1703	0.1643
cosine_precision@10	0.0899	0.0904	0.0886	0.0886	0.0873
cosine_recall@1	0.68	0.6686	0.6771	0.6543	0.6157
cosine_recall@3	0.8243	0.82	0.8143	0.8071	0.7771
cosine_recall@5	0.8571	0.86	0.8571	0.8514	0.8214
cosine_recall@10	0.8986	0.9043	0.8857	0.8857	0.8729
cosine_ndcg@10	0.7931	0.7907	0.787	0.7751	0.7473
cosine_mrr@10	0.759	0.754	0.7548	0.7391	0.7068
cosine_map@100	0.7628	0.7573	0.7593	0.743	0.7111

Training Details

Training Dataset

json

Dataset: json
Size: 6,300 training samples
Columns: positive and anchor
Approximate statistics based on the first 1000 samples:
positive anchor
type string string
details
min: 8 tokens
mean: 46.61 tokens
max: 439 tokens

min: 7 tokens
mean: 20.72 tokens
max: 51 tokens

	positive	anchor
type	string	string
details	min: 8 tokens mean: 46.61 tokens max: 439 tokens	min: 7 tokens mean: 20.72 tokens max: 51 tokens

Samples:

positive	anchor
`Operating costs and expenses increased $80.3 million, or 7.1%, during the year ended December 31, 2023, compared to the year ended December 31, 2022 primarily due to increases in film exhibition and food and beverage costs.`	`What factors contributed to the escalation in operating costs and expenses in 2023?`
`In the United States, the company purchases HFCS to meet its and its bottlers’ requirements with the assistance of Coca-Cola Bottlers’ Sales & Services Company LLC, which is a procurement service provider for their North American operations.`	`How does the company source high fructose corn syrup (HFCS) in the United States?`
`Item 8. Financial Statements and Supplementary Data The index to Financial Statements and Supplementary Data is presented`	`What is presented in Item 8 according to Financial Statements and Supplementary Data?`

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: 16
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: 16
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
0.4061	10	16.0873	-	-	-	-	-
0.8122	20	8.3282	-	-	-	-	-
1.0	25	-	0.7841	0.7796	0.7774	0.7631	0.7320
1.2030	30	5.1781	-	-	-	-	-
1.6091	40	4.0947	-	-	-	-	-
2.0	50	3.9824	0.7888	0.7867	0.7851	0.7701	0.7401
2.4061	60	2.854	-	-	-	-	-
2.8122	70	2.9878	-	-	-	-	-
3.0	75	-	0.7913	0.7903	0.7869	0.7755	0.7469
3.2030	80	2.5653	-	-	-	-	-
3.6091	90	2.999	-	-	-	-	-
3.8528	96	-	0.7931	0.7907	0.7870	0.7751	0.7473

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: 3.3.2
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}
}