metadata

language:
  - en
license: apache-2.0
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:185814
  - loss:MatryoshkaLoss
  - loss:MultipleNegativesRankingLoss
base_model: BAAI/bge-base-en-v1.5
widget:
  - source_sentence: ' The passage suggests that the more utility an item has, the more value human beings assign to it, making utility synonymous with subjective human value.'
    sentences:
      - |2
         In the given passage about compound interest, how does the interest earned on a Series EE bond affect its value over time? 
      - |2
         What forms does Section 16 require insiders to file and when is Form 3 typically submitted? 
      - |2
         What is the relationship between an item's utility and its subjective human value according to the passage? 
  - source_sentence: ' The price per share is determined when a company goes public by giving a valuation to the company with the input of an investment bank. This value is then divided by the total number of shares to be issued.'
    sentences:
      - |2
         How is the price per share determined when a company goes public and involves an investment bank? 
      - |2
         What percentage decrease have Fisker shares experienced in the past year despite a 30% increase on Feb. 27?
      - |2
         What factors contributed to the strong performance and outstanding returns of residential construction stocks since the March lows as mentioned in the passage? 
  - source_sentence: ' Municipal bonds are discussed as this type of investment.'
    sentences:
      - |2
         What is the benefit and process of using an income-driven repayment (IDR) plan for federal student loans? 
      - |2
         Which luxury watch and jewelry brands are owned by Swatch?
      - |2
         What type of investment is discussed as a way to potentially increase after-tax returns by avoiding federal taxes, and is often chosen for its relative safety and steady return? 
  - source_sentence: ' The rally could potentially fill the Sept. 18 gap between $145 and $150, reaching the .618 sell-off retracement level. The on-balance volume indicator suggests that Roku stock is unlikely to test the September high at this time.'
    sentences:
      - |2
         What rewards and perks can Navy Federal Visa Signature Flagship Rewards credit card users receive upon opening an account and within the first 90 days? 
      - |2
         What aspects contribute to VeriFirst's high marks for comprehensiveness in their services? 
      - |2
         What levels could the rally potentially fill after buyers buy the dip into $120, and what does the on-balance volume indicator suggest about Roku stock's likelihood to test the September high at this time? 
  - source_sentence: ' Home Depot''s stock closed at $135.39 while being above a "golden cross" on January 19, 2017.'
    sentences:
      - |2
         In the given text passage, when did Home Depot's stock close at $135.39 while being above a "golden cross"? 
      - |2
         What term does JPMorgan use to refer to net interest margin in its financial materials, and what was their net interest margin in FY 2019 before the pandemic started? 
      - |2
         According to Maley, where might the funds from potentially declining sectors like FANGs be directed towards? 
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: Regulatory Financial Matryoshka
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 768
          type: dim_768
        metrics:
          - type: cosine_accuracy@1
            value: 0.6027025718021989
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7349251707269822
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.7675691383736136
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8058313556448878
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.6027025718021989
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.24497505690899404
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.1535138276747227
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.0805831355644888
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.6027025718021989
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7349251707269822
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.7675691383736136
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8058313556448878
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.7073258915973659
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6754839282543154
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6796950515367028
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 512
          type: dim_512
        metrics:
          - type: cosine_accuracy@1
            value: 0.5988763500750715
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7302271516443066
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.7651474790526469
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.8033612631375018
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.5988763500750715
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.2434090505481022
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.15302949581052935
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.08033612631375019
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.5988763500750715
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7302271516443066
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.7651474790526469
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.8033612631375018
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.703957174859045
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6718512470776807
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6760676798344978
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 256
          type: dim_256
        metrics:
          - type: cosine_accuracy@1
            value: 0.5881241826899791
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7223325422579552
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.756768537802102
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7946917227684409
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.5881241826899791
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.24077751408598502
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.1513537075604204
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.07946917227684411
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.5881241826899791
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7223325422579552
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.756768537802102
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7946917227684409
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.694546619247058
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6621607466706108
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6665568671650335
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 128
          type: dim_128
        metrics:
          - type: cosine_accuracy@1
            value: 0.5718990652395021
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.7057683925025428
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.7405918535380442
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7818569283673172
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.5718990652395021
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.2352561308341809
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.14811837070760883
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.07818569283673171
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.5718990652395021
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.7057683925025428
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.7405918535380442
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7818569283673172
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6793129712551184
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6462535008352889
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6508176832915368
            name: Cosine Map@100
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: dim 64
          type: dim_64
        metrics:
          - type: cosine_accuracy@1
            value: 0.5443405821669007
            name: Cosine Accuracy@1
          - type: cosine_accuracy@3
            value: 0.6759335496682327
            name: Cosine Accuracy@3
          - type: cosine_accuracy@5
            value: 0.7138567346345716
            name: Cosine Accuracy@5
          - type: cosine_accuracy@10
            value: 0.7583183997675207
            name: Cosine Accuracy@10
          - type: cosine_precision@1
            value: 0.5443405821669007
            name: Cosine Precision@1
          - type: cosine_precision@3
            value: 0.22531118322274424
            name: Cosine Precision@3
          - type: cosine_precision@5
            value: 0.14277134692691432
            name: Cosine Precision@5
          - type: cosine_precision@10
            value: 0.07583183997675207
            name: Cosine Precision@10
          - type: cosine_recall@1
            value: 0.5443405821669007
            name: Cosine Recall@1
          - type: cosine_recall@3
            value: 0.6759335496682327
            name: Cosine Recall@3
          - type: cosine_recall@5
            value: 0.7138567346345716
            name: Cosine Recall@5
          - type: cosine_recall@10
            value: 0.7583183997675207
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6522706632460163
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6182239473662035
            name: Cosine Mrr@10
          - type: cosine_map@100
            value: 0.6229041572175256
            name: Cosine Map@100

Regulatory 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("hshashank06/regulatory-model")
# Run inference
sentences = [
    ' Home Depot\'s stock closed at $135.39 while being above a "golden cross" on January 19, 2017.',
    ' In the given text passage, when did Home Depot\'s stock close at $135.39 while being above a "golden cross"? \n',
    ' According to Maley, where might the funds from potentially declining sectors like FANGs be directed towards? \n',
]
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	0.6027	0.5989	0.5881	0.5719	0.5443
cosine_accuracy@3	0.7349	0.7302	0.7223	0.7058	0.6759
cosine_accuracy@5	0.7676	0.7651	0.7568	0.7406	0.7139
cosine_accuracy@10	0.8058	0.8034	0.7947	0.7819	0.7583
cosine_precision@1	0.6027	0.5989	0.5881	0.5719	0.5443
cosine_precision@3	0.245	0.2434	0.2408	0.2353	0.2253
cosine_precision@5	0.1535	0.153	0.1514	0.1481	0.1428
cosine_precision@10	0.0806	0.0803	0.0795	0.0782	0.0758
cosine_recall@1	0.6027	0.5989	0.5881	0.5719	0.5443
cosine_recall@3	0.7349	0.7302	0.7223	0.7058	0.6759
cosine_recall@5	0.7676	0.7651	0.7568	0.7406	0.7139
cosine_recall@10	0.8058	0.8034	0.7947	0.7819	0.7583
cosine_ndcg@10	0.7073	0.704	0.6945	0.6793	0.6523
cosine_mrr@10	0.6755	0.6719	0.6622	0.6463	0.6182
cosine_map@100	0.6797	0.6761	0.6666	0.6508	0.6229

Training Details

Training Dataset

json

Dataset: json
Size: 185,814 training samples
Columns: positive and anchor
Approximate statistics based on the first 1000 samples:
positive anchor
type string string
details
min: 3 tokens
mean: 43.18 tokens
max: 200 tokens

min: 10 tokens
mean: 23.08 tokens
max: 63 tokens

	positive	anchor
type	string	string
details	min: 3 tokens mean: 43.18 tokens max: 200 tokens	min: 10 tokens mean: 23.08 tokens max: 63 tokens

Samples:

positive	anchor
`The BVPS (Book Value Per Share) is calculated by dividing a company's common equity value by its total number of shares outstanding. In the given example, if a company has a common equity value of $100 million and 10 million shares outstanding, its BVPS would be $10 ($100 million / 10 million). You can calculate a company's BVPS using Microsoft Excel by entering the values of common stock, retained earnings, and additional paid-in capital into cells A1 through A3.`	`What is the BVPS and how is it calculated?`
`They facilitate commodities trading using their resources, can take delivery of commodities if needed, provide advisory services for clients, and act as market makers by buying and selling futures contracts to add liquidity to the marketplace. The passage uses the example of a commercial baking firm to demonstrate how their impact can be seen in the market.`	`What role do eligible commercial entities play in commodities trading and market liquidity?`
`Naive diversification is a type of diversification strategy where an investor randomly selects different securities, hoping to lower the risk of the portfolio due to the varied nature of the chosen securities. It is less sophisticated than diversification methods using statistical modeling, but when guided by experience, careful security examination, and common sense, it remains an effective strategy for reducing portfolio risk.`	`What is the concept of naive diversification in investing and how does it compare to more sophisticated diversification methods?`

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
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: 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: 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

Click to expand

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
0.0276	10	43.573	-	-	-	-	-
0.0551	20	42.1758	-	-	-	-	-
0.0827	30	37.6368	-	-	-	-	-
0.1102	40	34.5743	-	-	-	-	-
0.1378	50	29.5956	-	-	-	-	-
0.1653	60	23.4468	-	-	-	-	-
0.1929	70	19.7425	-	-	-	-	-
0.2204	80	16.9744	-	-	-	-	-
0.2480	90	15.2437	-	-	-	-	-
0.2755	100	13.9444	-	-	-	-	-
0.3031	110	12.067	-	-	-	-	-
0.3306	120	11.1149	-	-	-	-	-
0.3582	130	10.4083	-	-	-	-	-
0.3857	140	8.915	-	-	-	-	-
0.4133	150	9.4964	-	-	-	-	-
0.4408	160	8.0434	-	-	-	-	-
0.4684	170	8.1963	-	-	-	-	-
0.4960	180	8.5704	-	-	-	-	-
0.5235	190	7.711	-	-	-	-	-
0.5511	200	7.6676	-	-	-	-	-
0.5786	210	6.9899	-	-	-	-	-
0.6062	220	7.6195	-	-	-	-	-
0.6337	230	7.0456	-	-	-	-	-
0.6613	240	7.5541	-	-	-	-	-
0.6888	250	6.6543	-	-	-	-	-
0.7164	260	6.8849	-	-	-	-	-
0.7439	270	7.6635	-	-	-	-	-
0.7715	280	7.2155	-	-	-	-	-
0.7990	290	6.3284	-	-	-	-	-
0.8266	300	6.577	-	-	-	-	-
0.8541	310	5.0835	-	-	-	-	-
0.8817	320	6.1866	-	-	-	-	-
0.9092	330	5.9467	-	-	-	-	-
0.9368	340	5.663	-	-	-	-	-
0.9644	350	5.417	-	-	-	-	-
0.9919	360	6.0331	-	-	-	-	-
0.9974	362	-	0.6940	0.6900	0.6791	0.6603	0.6273
1.0220	370	5.5374	-	-	-	-	-
1.0496	380	4.5917	-	-	-	-	-
1.0771	390	4.6483	-	-	-	-	-
1.1047	400	4.96	-	-	-	-	-
1.1323	410	4.6808	-	-	-	-	-
1.1598	420	5.2396	-	-	-	-	-
1.1874	430	4.651	-	-	-	-	-
1.2149	440	4.4875	-	-	-	-	-
1.2425	450	4.6877	-	-	-	-	-
1.2700	460	4.2209	-	-	-	-	-
1.2976	470	4.678	-	-	-	-	-
1.3251	480	4.6774	-	-	-	-	-
1.3527	490	4.4409	-	-	-	-	-
1.3802	500	4.4464	-	-	-	-	-
1.4078	510	4.2724	-	-	-	-	-
1.4353	520	4.5017	-	-	-	-	-
1.4629	530	4.3469	-	-	-	-	-
1.4904	540	4.4925	-	-	-	-	-
1.5180	550	3.922	-	-	-	-	-
1.5455	560	4.6949	-	-	-	-	-
1.5731	570	4.0364	-	-	-	-	-
1.6007	580	4.3846	-	-	-	-	-
1.6282	590	3.7526	-	-	-	-	-
1.6558	600	4.0508	-	-	-	-	-
1.6833	610	4.6315	-	-	-	-	-
1.7109	620	3.7683	-	-	-	-	-
1.7384	630	4.6994	-	-	-	-	-
1.7660	640	4.1994	-	-	-	-	-
1.7935	650	4.3915	-	-	-	-	-
1.8211	660	4.2947	-	-	-	-	-
1.8486	670	4.6972	-	-	-	-	-
1.8762	680	4.1664	-	-	-	-	-
1.9037	690	4.1861	-	-	-	-	-
1.9313	700	3.6879	-	-	-	-	-
1.9588	710	4.3767	-	-	-	-	-
1.9864	720	4.48	-	-	-	-	-
1.9974	724	-	0.7013	0.6971	0.6885	0.6716	0.6414
2.0165	730	3.6164	-	-	-	-	-
2.0441	740	3.3361	-	-	-	-	-
2.0716	750	3.4175	-	-	-	-	-
2.0992	760	3.9006	-	-	-	-	-
2.1267	770	3.0823	-	-	-	-	-
2.1543	780	3.029	-	-	-	-	-
2.1818	790	3.8081	-	-	-	-	-
2.2094	800	3.4486	-	-	-	-	-
2.2370	810	3.6064	-	-	-	-	-
2.2645	820	3.0896	-	-	-	-	-
2.2921	830	3.3233	-	-	-	-	-
2.3196	840	2.9528	-	-	-	-	-
2.3472	850	3.0482	-	-	-	-	-
2.3747	860	3.2795	-	-	-	-	-
2.4023	870	2.9218	-	-	-	-	-
2.4298	880	3.4518	-	-	-	-	-
2.4574	890	3.6095	-	-	-	-	-
2.4849	900	3.2002	-	-	-	-	-
2.5125	910	3.368	-	-	-	-	-
2.5400	920	3.0623	-	-	-	-	-
2.5676	930	3.3495	-	-	-	-	-
2.5951	940	3.7123	-	-	-	-	-
2.6227	950	3.7795	-	-	-	-	-
2.6502	960	3.5567	-	-	-	-	-
2.6778	970	3.3498	-	-	-	-	-
2.7054	980	3.3141	-	-	-	-	-
2.7329	990	2.9425	-	-	-	-	-
2.7605	1000	2.9978	-	-	-	-	-
2.7880	1010	3.2468	-	-	-	-	-
2.8156	1020	2.5252	-	-	-	-	-
2.8431	1030	3.3108	-	-	-	-	-
2.8707	1040	3.195	-	-	-	-	-
2.8982	1050	3.1019	-	-	-	-	-
2.9258	1060	3.7059	-	-	-	-	-
2.9533	1070	3.1952	-	-	-	-	-
2.9809	1080	3.2454	-	-	-	-	-
2.9974	1086	-	0.7056	0.7030	0.6939	0.6779	0.6505
3.0110	1090	3.3788	-	-	-	-	-
3.0386	1100	2.9617	-	-	-	-	-
3.0661	1110	3.4313	-	-	-	-	-
3.0937	1120	2.5883	-	-	-	-	-
3.1212	1130	2.8836	-	-	-	-	-
3.1488	1140	2.3895	-	-	-	-	-
3.1763	1150	2.5155	-	-	-	-	-
3.2039	1160	3.3168	-	-	-	-	-
3.2314	1170	3.0286	-	-	-	-	-
3.2590	1180	3.1494	-	-	-	-	-
3.2866	1190	2.87	-	-	-	-	-
3.3141	1200	2.591	-	-	-	-	-
3.3417	1210	2.8437	-	-	-	-	-
3.3692	1220	3.0344	-	-	-	-	-
3.3968	1230	3.0685	-	-	-	-	-
3.4243	1240	3.4623	-	-	-	-	-
3.4519	1250	3.4256	-	-	-	-	-
3.4794	1260	2.7349	-	-	-	-	-
3.5070	1270	2.8587	-	-	-	-	-
3.5345	1280	2.729	-	-	-	-	-
3.5621	1290	3.0288	-	-	-	-	-
3.5896	1300	2.6599	-	-	-	-	-
3.6172	1310	2.4755	-	-	-	-	-
3.6447	1320	3.0501	-	-	-	-	-
3.6723	1330	2.545	-	-	-	-	-
3.6998	1340	2.5919	-	-	-	-	-
3.7274	1350	2.9026	-	-	-	-	-
3.7550	1360	2.7362	-	-	-	-	-
3.7825	1370	3.3311	-	-	-	-	-
3.8101	1380	2.8415	-	-	-	-	-
3.8376	1390	3.2033	-	-	-	-	-
3.8652	1400	2.7483	-	-	-	-	-
3.8927	1410	3.0403	-	-	-	-	-
3.9203	1420	3.0724	-	-	-	-	-
3.9478	1430	2.9797	-	-	-	-	-
3.9754	1440	2.6779	-	-	-	-	-
3.9974	1448	-	0.7073	0.704	0.6945	0.6793	0.6523

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.11.11
Sentence Transformers: 3.4.1
Transformers: 4.48.3
PyTorch: 2.6.0+cu124
Accelerate: 1.3.0
Datasets: 3.4.1
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}
}