Hugging Face's logo

icapital-alpha-user
/
stella_400m_trained_v0

Sentence Similarity
sentence-transformers
Safetensors
new
feature-extraction
Generated from Trainer
dataset_size:43536
loss:CosineSimilarityLoss
custom_code
text-embeddings-inference
Inference Endpoints
Model card Files Community
Edit model card

SentenceTransformer

This is a sentence-transformers model trained on the json dataset. It maps sentences & paragraphs to a 2048-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
  • Maximum Sequence Length: 2048 tokens
  • Output Dimensionality: 2048 dimensions
  • Similarity Function: Cosine Similarity
  • Training Dataset:
    • json

Model Sources

Full Model Architecture 

SentenceTransformer(
  (0): Transformer({'max_seq_length': 2048, 'do_lower_case': False}) with Transformer model: NewModel 
  (1): Pooling({'word_embedding_dimension': 1024, '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): Dense({'in_features': 1024, 'out_features': 2048, 'bias': True, 'activation_function': 'torch.nn.modules.linear.Identity'})
)

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("icapital-alpha-user/stella_400m_trained_v0")
# Run inference
sentences = [
    '\n        ### The Question:\n        What is the tax structure?\n        \n        ### Definition:\n        The tax classification of the product where it is domiciled that determines how it is treated for federal income tax purposes.\n\n        ',
    'PPM2022102563.1 : iCapital\r\n- 19 - \r\n1.6.175 “Valuer” means Goldman Sachs & Co. LLC, as appointed by the AIFM as its \r\ndelegate, to provide valuation services in relation to the assets of the Partnership. \r\n1.6.176 “VAT” means: (a) any tax imposed in compliance with the council directive of 28 \r\nNovember 2006 on the common system of value added tax (EC Directive 2006/112), and (b) any \r\nother tax of a similar nature, whether imposed in a member state of the European Union in \r\nsubstitution for, or levied in addition to, such tax referred to in (a) above, or elsewhere (including \r\nconsumption tax, goods and services tax, sales tax and turnover tax). \r\n1.6.177 “Vintage Funds” means the private funds managed by the Investment Advisor \r\nthat focus primarily on making secondary investments in funds. \r\n1.6.178 “Vintage IX A Funds” means, collectively, (i) the Delaware Vintage IX A \r\nPartnership, and (ii) the Luxembourg Vintage IX A Partnership. \r\n1.6.179 “Vintage IX B Funds” means, collectively, (i) the Partnership, and (ii) the \r\nDelaware Partnership. \r\n1.6.180 “Vintage IX Employee Fund” means an Employee Fund that is formed to invest \r\nall or a designated portion of its capital alongside a Vintage IX Fund intended for external investors \r\n(or the portion of its capital to be invested alongside a Vintage IX Fund if only a designated portion \r\nof its capital will be invested alongside the other Vintage IX Funds). \r\n1.6.181 “Vintage IX Funds” means, collectively, (i) the Vintage IX A Funds, (ii) the \r\nVintage IX B Funds, (iii) the Vintage IX Employee Funds, and (iv) any other parallel or feeder fund \r\nformed by the General Partner or the Investment Advisor which will invest substantially all of its \r\nassets directly or indirectly in or alongside the Partnership and the Delaware Partnership and is \r\ndesignated as a “Vintage IX Fund” for such purposes by the Investment Advisor in its discretion. \r\n1.6.182 “Vintage Team” means the team within the Investment Advisor that manages \r\nthe Vintage Funds. \r\n1.6.183 “Voting Matter” is defined in Section 12.1.3. \r\nFor all purposes of this Agreement and any schedules and exhibits hereto, except as \r\nexpressly provided herein or unless the context otherwise requires, the words “including,” \r\n“includes,” “include” and words of similar import shall be deemed to be followed by the phrase \r\n“without limitation” and shall be regarded as a reference to non-exclusive and non-characterizing \r\nillustrations. Except as otherwise expressly provided herein, in any case where Goldman Sachs, \r\nthe General Partner, the AIFM, the Investment Manager or the Investment Advisor is authorized or \r\nrequired to take an action, exercise its discretion, make any determination or give any approval, it \r\nshall do so in its sole discretion or sole judgment taking into account any considerations it deems \r\nappropriate. It is intended that the terms of this Agreement be construed in accordance with their \r\nfair meanings and not against any particular Person, including the General Partner, the AIFM, the \r\nInvestment Manager or the Investment Advisor. Except as otherwise provided herein, all non-U.S. \r\ndollar assets and liabilities shall be considered to be the U.S. dollar equivalent thereof at the then \r\napplicable conversion spot rates as determined by the General Partner. References to the \r\n“General Partner” herein shall mean the General Partner acting as managing general partner \r\n(associé gérant commandité) of the Partnership or its duly appointed delegates, as applicable. \r\niCapital Network - 10.30.2024 [email protected]\n',
    'Proprietary and Confidential\r\n7\r\ncapital contribution to the Access Fund, Lead Edge or the Underlying Fund\r\nmay impose certain remedies against the Access Fund, including without \r\nlimitation, causing the Access Fund to forfeit all or a portion of its interest in \r\nthe Underlying Fund.\r\nWith respect to any capital contribution (or portion thereof) that is subject to \r\na default (the “Defaulted Amount”), the General Partner may, in consultation \r\nwith the Investment Advisor, call additional capital from the Limited Partners \r\nthat have already made the applicable capital contribution (not in excess of \r\ntheir unfunded Subscriptions) to the extent necessary to fund the Defaulted \r\nAmount upon five (5) days written notice (or such shorter amount of time as \r\nwas required for the initial capital contributions that required the deficiency \r\ndrawdown).\r\nAs described in detail in the Partnership Agreement, the General Partner has \r\nagreed to give the Investment Advisor certain consent or consultation rights \r\nwith respect to defaults on capital contributions.\r\nBifurcated Default If the Access Fund fails to contribute all or any portion of any capital call \r\namount set forth in a funding notice received from the Underlying Fund as \r\nand when due, and such failure results from the failure of one or more Limited \r\nPartners or limited partners of a Feeder Fund (each such Limited Partner or \r\nlimited partner of a Feeder Fund, as applicable, a “Defaulting Access Fund \r\nPartner”) to make full payment in respect of any capital call issued by the \r\nAccess Fund, then the Lead Edge GP has agreed to only treat the Access Fund \r\nas a “Defaulting Partner” (as such term is defined in the Underlying Fund \r\nAgreement) with respect to the portion of the Access Fund’s interest in the \r\nUnderlying Fund that has defaulted.\r\nDistributions Distributions from the Underlying Fund received by the Access Fund will \r\ngenerally be distributed to the Limited Partners pro rata based on their \r\nrespective Subscriptions in the Access Fund as promptly as practicable, \r\nsubject to the Investment Management Fee, Management Fee and Tax \r\nReporting Oversight Services Fee. In addition, the Access Fund will be \r\nentitled to withhold from any distribution amounts necessary to create, in the \r\nGeneral Partner’s sole discretion, but in consultation with the Investment \r\nAdvisor, reasonable reserves for the payment of Access Fund expenses and \r\nliabilities, to make anticipated capital contributions to the Underlying Fund\r\nand for any other purpose permitted under the Partnership Agreement.\r\nLiquidating distributions will be made in accordance with positive capital \r\naccount balances.\r\nAs described in detail in the Partnership Agreement, the General Partner has \r\nagreed to give the Investment Advisor certain consent or consultation rights \r\nwith respect to distributions.\r\niCapital Network - 10.30.2024 [email protected]\n',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 2048]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Training Details

Training Dataset

json

  • Dataset: json
  • Size: 43,536 training samples
  • Columns: question, context, and score
  • Approximate statistics based on the first 1000 samples:
    question context score
    type string string float
    details
    • min: 26 tokens
    • mean: 47.33 tokens
    • max: 159 tokens
    • min: 54 tokens
    • mean: 730.39 tokens
    • max: 2048 tokens
    • min: 0.0
    • mean: 0.26
    • max: 1.0
  • Samples:
    question context score

    ### Question
    What is the currency applies to this fund?
    ### Definition
    "The currency the fund is denominated in.

    		 Proprietary and Confidential

    45

    of certain earnings and profits of the CFC attributable to such stock. Further, if a Limited Partner disposes

    of its Interests, the Limited Partner may be required to recognize ordinary income under Section 751 of the

    Code equal to its distributive share of any Section 1248 income that would have been triggered if the

    Underlying Fund had sold its interest in the CFC at fair market value. There may also be a tax imposed on

    a U.S. Shareholder’s aggregate net CFC income that is treated as “global intangible low-taxed income.” As

    a result of the foregoing, a Limited Partner may be required to recognize income sooner than it otherwise

    would. There can be no assurance that the Access Fund will be aware if it invests in a CFC or whether it

    is a U.S. Shareholder for these purposes.

    The application of the CFC rules is complex, and may require Limited Partners to file certain forms with

    the IRS. The Access Fund is under no obligation to assist a ...    		 0.0

    ### The Question:
    What is the tax structure?

    ### Definition:
    The tax classification of the product where it is domiciled that determines how it is treated for federal income tax purposes.

    		 Certain Legal and Tax Considerations

    Strategic Partners IX 191

    an Underlying Vehicle as a PFIC or a CFC may result in adverse tax consequences for Limited Partners. In

    addition, if an Underlying Vehicle were treated as a non-U.S. corporation, it would be subject, at the entity

    level, to the U.S. federal income tax consequences outlined in “—Non-U.S. Investors in the Partnership”

    below. If the Partnership or any Underlying Vehicle that is organized in the United States were treated as

    a corporation for U.S. federal income tax purposes, the Partnership or the relevant Underlying Vehicle

    would generally be subject to U.S. federal income tax (currently, at a rate of 21%) on all of its net income,

    as well as any applicable state and local taxes, and certain distributions that it made, other than in

    liquidation, generally would be taxable to its investors as dividends to the extent of its current and

    accumulated earnings and profits.

    As partnerships, neither the Partnershi...    		 0.0

    ### Question
    What is the investor qualification standard applied to this fund?
    ### Definition
    The qualification standard investors must meet in order to be eligible to invest in the fund.

    		 Proprietary and Confidential

    46

    Each investor must be prepared to bear the economic risk of the investment in the Interests for an indefinite

    period because the Interests cannot be sold unless they are subsequently registered under the Securities Act

    or an exemption for such registration is available. It is extremely unlikely that the Interests will ever be

    registered under the Securities Act. The Interests may not be transferred or resold except as permitted under

    the Securities Act and any other applicable securities laws, pursuant to registration or exemption therefrom.

    As described elsewhere in this Memorandum, the transferability of the Interests will be further restricted

    by the terms of the Partnership Agreement.

    Securities Exchange Act of 1934. It is not expected that the Access Fund will be required to register the

    Interests under Section 12(g) or any other provision of the Exchange Act. The Access Fund does not expect

    to have any class of equity security he...    		 0.0
  • Loss: CosineSimilarityLoss with these parameters:
    {
    "loss_fct": "torch.nn.modules.loss.MSELoss"
}

Evaluation Dataset

json

  • Dataset: json
  • Size: 43,536 evaluation samples
  • Columns: question, context, and score
  • Approximate statistics based on the first 1000 samples:
    question context score
    type string string float
    details
    • min: 26 tokens
    • mean: 47.11 tokens
    • max: 159 tokens
    • min: 47 tokens
    • mean: 744.25 tokens
    • max: 2048 tokens
    • min: 0.0
    • mean: 0.26
    • max: 1.0
  • Samples:
    question context score

    ### Question
    What is the jurisdiction applies to this fund?
    ### Definition
    The physical location where the fund is incorporated

    		 1

    I. INTRODUCTION

    iCapital-PCRED II Onshore Access Fund, L.P., a Delaware limited partnership (the “Access Fund”), has

    been formed to invest substantially all of its investable assets in PIMCO Commercial Real Estate Debt Fund

    II Special Onshore Feeder, L.P., a Delaware limited partnership (together with any parallel funds, master

    funds, feeder vehicles or alternative investment vehicles, the “Underlying Fund”). The principal investment

    objective of the Underlying Fund is as described in the Underlying Fund PPM and generally is to seek to

    provide investors with attractive returns through investments through one or more U.S. intermediate

    investment vehicles electing to be treated as real estate investment trusts for U.S. federal income tax

    purposes, which intermediate investment vehicles will in turn invest in the Master Fund. The Master Fund

    will seek to originate and/or acquire predominantly U.S. (i) senior commercial real estate (“CRE”) first

    mortgage loans on CRE ...    		 1.0

    ### Question
    What is the jurisdiction applies to this fund?
    ### Definition
    The physical location where the fund is incorporated

    		 Proprietary and Confidential

    50

    Underlying Fund and the Access Fund are subject. Each prospective investor should consider the potential

    consequences of such taxes on an investment in the Access Fund. It is the responsibility of each prospective

    investor to satisfy itself as to, among other things, the legal and tax consequences of an investment in the

    Access Fund under state law, including the laws of the state(s) of its domicile and its residence, by obtaining

    advice from its own tax advisor, and to file all appropriate tax returns that may be required.

    The Access Fund may also bear substantial non-U.S. taxes as well.

    The foregoing is a summary of some of the important U.S. federal income tax rules and considerations

    affecting the Limited Partners and the Access Fund’s operations and does not purport to be a

    complete analysis of all relevant tax rules and considerations, nor does it purport to be a complete

    listing of all potential tax risks inherent in purchasing o...    		 0.0

    ### Question
    What is the Fiscal Year End applies to this fund?
    ### Definition
    "The date that is considered as end of a fiscal year for calculating annual financial statements. It should be one of the four possible dates [30-Jun, 30-Sep, 31-Mar, 31-Dec]. If the document does not mention specific date but clearly mention that the fiscal year is calendar year then the fiscal year will be Dec 31"

    		 ix

    THIS MEMORANDUM CONTAINS FORWARD-LOOKING STATEMENTS, WHICH CAN BE

    IDENTIFIED BY THE USE OF FORWARD-LOOKING TERMINOLOGY SUCH AS “MAY,” “SEEK,”

    “EXPECT,” “ESTIMATE,” OR “BELIEVE,” OR THE NEGATIVES THEREOF OR OTHER

    VARIATIONS THEREON OR COMPARABLE TERMINOLOGY. FORWARD-LOOKING

    STATEMENTS ARE STATEMENTS THAT ARE NOT HISTORICAL FACTS, INCLUDING

    STATEMENTS ABOUT BELIEFS AND EXPECTATIONS, AND THE USE OF QUALITATIVE

    TERMS SUCH AS “BEST-OF-CLASS,” “SUPERIOR,” AND “TOP-TIER.” ANY STATEMENT IN

    THIS MEMORANDUM THAT CONTAINS INTENTIONS, BELIEFS, EXPECTATIONS, OR

    PREDICTIONS (AND THE ASSUMPTIONS UNDERLYING THEM) IS A FORWARD-LOOKING

    STATEMENT. THESE STATEMENTS ARE BASED ON PLANS, ESTIMATES, AND PROJECTIONS,

    AS THEY ARE CURRENTLY AVAILABLE. FORWARD-LOOKING STATEMENTS THEREFORE

    SPEAK ONLY AS OF THE DATE THEY ARE MADE, AND NONE OF THE PLACEMENT AGENTS

    (IF ANY), THE FEEDER FUNDS, THE FEEDER FUND GENERAL PARTNERS, OR THE

    INVESTMENT MANAGER UNDERTAKE TO UPDATE ANY OF THEM IN L...    		 0.0
  • Loss: CosineSimilarityLoss with these parameters:
    {
    "loss_fct": "torch.nn.modules.loss.MSELoss"
}

Training Hyperparameters

Non-Default Hyperparameters

  • eval_strategy: steps
  • per_device_train_batch_size: 4
  • per_device_eval_batch_size: 32
  • learning_rate: 2e-05
  • num_train_epochs: 2

All Hyperparameters

Click to expand
  • overwrite_output_dir: False
  • do_predict: False
  • eval_strategy: steps
  • prediction_loss_only: True
  • per_device_train_batch_size: 4
  • per_device_eval_batch_size: 32
  • 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: 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: 2
  • 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: False
  • 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: proportional

Training Logs

Click to expand
Epoch Step Training Loss Validation Loss
0.0051 50 0.1587 -
0.0102 100 0.1123 -
0.0153 150 0.1421 -
0.0204 200 0.1193 -
0.0255 250 0.0967 -
0.0306 300 0.0998 -
0.0357 350 0.072 -
0.0408 400 0.124 -
0.0459 450 0.1076 -
0.0510 500 0.0738 -
0.0561 550 0.0727 -
0.0612 600 0.0794 -
0.0664 650 0.0821 -
0.0715 700 0.0712 -
0.0766 750 0.0629 -
0.0817 800 0.0874 -
0.0868 850 0.0552 -
0.0919 900 0.1073 -
0.0970 950 0.0713 -
0.1021 1000 0.0554 0.0622
0.1072 1050 0.0839 -
0.1123 1100 0.0515 -
0.1174 1150 0.0775 -
0.1225 1200 0.0661 -
0.1276 1250 0.0737 -
0.1327 1300 0.0863 -
0.1378 1350 0.0563 -
0.1429 1400 0.0592 -
0.1480 1450 0.0724 -
0.1531 1500 0.0632 -
0.1582 1550 0.0612 -
0.1633 1600 0.0731 -
0.1684 1650 0.0435 -
0.1735 1700 0.0764 -
0.1786 1750 0.0591 -
0.1837 1800 0.0498 -
0.1889 1850 0.0689 -
0.1940 1900 0.0557 -
0.1991 1950 0.0519 -
0.2042 2000 0.0727 0.0600
0.2093 2050 0.0475 -
0.2144 2100 0.053 -
0.2195 2150 0.0583 -
0.2246 2200 0.0361 -
0.2297 2250 0.0526 -
0.2348 2300 0.0371 -
0.2399 2350 0.0722 -
0.2450 2400 0.0671 -
0.2501 2450 0.0409 -
0.2552 2500 0.053 -
0.2603 2550 0.0412 -
0.2654 2600 0.0602 -
0.2705 2650 0.0583 -
0.2756 2700 0.0475 -
0.2807 2750 0.0732 -
0.2858 2800 0.0484 -
0.2909 2850 0.0685 -
0.2960 2900 0.0575 -
0.3011 2950 0.0604 -
0.3062 3000 0.0815 0.0472
0.3114 3050 0.0635 -
0.3165 3100 0.0536 -
0.3216 3150 0.0594 -
0.3267 3200 0.0431 -
0.3318 3250 0.03 -
0.3369 3300 0.0723 -
0.3420 3350 0.0366 -
0.3471 3400 0.0623 -
0.3522 3450 0.0329 -
0.3573 3500 0.0363 -
0.3624 3550 0.049 -
0.3675 3600 0.0725 -
0.3726 3650 0.0374 -
0.3777 3700 0.0768 -
0.3828 3750 0.0556 -
0.3879 3800 0.0622 -
0.3930 3850 0.0415 -
0.3981 3900 0.0638 -
0.4032 3950 0.026 -
0.4083 4000 0.0601 0.0523
0.4134 4050 0.0804 -
0.4185 4100 0.0416 -
0.4236 4150 0.0479 -
0.4287 4200 0.0371 -
0.4339 4250 0.0709 -
0.4390 4300 0.0307 -
0.4441 4350 0.0463 -
0.4492 4400 0.0343 -
0.4543 4450 0.0385 -
0.4594 4500 0.0253 -
0.4645 4550 0.0405 -
0.4696 4600 0.024 -
0.4747 4650 0.0511 -
0.4798 4700 0.0287 -
0.4849 4750 0.0405 -
0.4900 4800 0.0394 -
0.4951 4850 0.0499 -
0.5002 4900 0.0517 -
0.5053 4950 0.0554 -
0.5104 5000 0.0496 0.0542
0.5155 5050 0.0586 -
0.5206 5100 0.0411 -
0.5257 5150 0.0478 -
0.5308 5200 0.0639 -
0.5359 5250 0.0435 -
0.5410 5300 0.034 -
0.5461 5350 0.0342 -
0.5512 5400 0.0392 -
0.5563 5450 0.0408 -
0.5615 5500 0.0434 -
0.5666 5550 0.0308 -
0.5717 5600 0.035 -
0.5768 5650 0.049 -
0.5819 5700 0.0411 -
0.5870 5750 0.0287 -
0.5921 5800 0.0339 -
0.5972 5850 0.0513 -
0.6023 5900 0.0385 -
0.6074 5950 0.029 -
0.6125 6000 0.059 0.0508
0.6176 6050 0.0439 -
0.6227 6100 0.049 -
0.6278 6150 0.042 -
0.6329 6200 0.053 -
0.6380 6250 0.0332 -
0.6431 6300 0.0236 -
0.6482 6350 0.0297 -
0.6533 6400 0.046 -
0.6584 6450 0.0514 -
0.6635 6500 0.0489 -
0.6686 6550 0.0299 -
0.6737 6600 0.0421 -
0.6788 6650 0.0346 -
0.6840 6700 0.0608 -
0.6891 6750 0.0305 -
0.6942 6800 0.0416 -
0.6993 6850 0.0518 -
0.7044 6900 0.0521 -
0.7095 6950 0.0184 -
0.7146 7000 0.0374 0.0406
0.7197 7050 0.0336 -
0.7248 7100 0.0382 -
0.7299 7150 0.056 -
0.7350 7200 0.0445 -
0.7401 7250 0.0537 -
0.7452 7300 0.043 -
0.7503 7350 0.0574 -
0.7554 7400 0.027 -
0.7605 7450 0.0396 -
0.7656 7500 0.0561 -
0.7707 7550 0.0311 -
0.7758 7600 0.0545 -
0.7809 7650 0.0483 -
0.7860 7700 0.0309 -
0.7911 7750 0.0511 -
0.7962 7800 0.0367 -
0.8013 7850 0.0364 -
0.8065 7900 0.0362 -
0.8116 7950 0.0419 -
0.8167 8000 0.0381 0.0423
0.8218 8050 0.0535 -
0.8269 8100 0.0441 -
0.8320 8150 0.0262 -
0.8371 8200 0.0251 -
0.8422 8250 0.0301 -
0.8473 8300 0.0247 -
0.8524 8350 0.0237 -
0.8575 8400 0.0478 -
0.8626 8450 0.0272 -
0.8677 8500 0.0569 -
0.8728 8550 0.047 -
0.8779 8600 0.0458 -
0.8830 8650 0.0392 -
0.8881 8700 0.0333 -
0.8932 8750 0.036 -
0.8983 8800 0.0272 -
0.9034 8850 0.034 -
0.9085 8900 0.0526 -
0.9136 8950 0.0263 -
0.9187 9000 0.0378 0.0421
0.9238 9050 0.0373 -
0.9290 9100 0.0262 -
0.9341 9150 0.0323 -
0.9392 9200 0.0175 -
0.9443 9250 0.0315 -
0.9494 9300 0.0349 -
0.9545 9350 0.0622 -
0.9596 9400 0.0245 -
0.9647 9450 0.0379 -
0.9698 9500 0.0486 -
0.9749 9550 0.0423 -
0.9800 9600 0.0479 -
0.9851 9650 0.0203 -
0.9902 9700 0.0201 -
0.9953 9750 0.048 -
1.0004 9800 0.0369 -
1.0055 9850 0.0358 -
1.0106 9900 0.0343 -
1.0157 9950 0.0493 -
1.0208 10000 0.0117 0.0345
1.0259 10050 0.0321 -
1.0310 10100 0.034 -
1.0361 10150 0.0316 -
1.0412 10200 0.0132 -
1.0463 10250 0.0149 -
1.0514 10300 0.0436 -
1.0566 10350 0.0428 -
1.0617 10400 0.0325 -
1.0668 10450 0.0348 -
1.0719 10500 0.0302 -
1.0770 10550 0.0276 -
1.0821 10600 0.0359 -
1.0872 10650 0.0186 -
1.0923 10700 0.0265 -
1.0974 10750 0.0231 -
1.1025 10800 0.0298 -
1.1076 10850 0.0394 -
1.1127 10900 0.0283 -
1.1178 10950 0.0345 -
1.1229 11000 0.0194 0.0365
1.1280 11050 0.0217 -
1.1331 11100 0.0208 -
1.1382 11150 0.0133 -
1.1433 11200 0.0099 -
1.1484 11250 0.0359 -
1.1535 11300 0.0089 -
1.1586 11350 0.0283 -
1.1637 11400 0.0439 -
1.1688 11450 0.0416 -
1.1739 11500 0.0179 -
1.1791 11550 0.0199 -
1.1842 11600 0.0415 -
1.1893 11650 0.0209 -
1.1944 11700 0.0218 -
1.1995 11750 0.0239 -
1.2046 11800 0.0243 -
1.2097 11850 0.0162 -
1.2148 11900 0.0411 -
1.2199 11950 0.0235 -
1.2250 12000 0.0389 0.0335
1.2301 12050 0.0411 -
1.2352 12100 0.0272 -
1.2403 12150 0.018 -
1.2454 12200 0.0268 -
1.2505 12250 0.0408 -
1.2556 12300 0.0389 -
1.2607 12350 0.0106 -
1.2658 12400 0.0287 -
1.2709 12450 0.0193 -
1.2760 12500 0.0308 -
1.2811 12550 0.0168 -
1.2862 12600 0.0466 -
1.2913 12650 0.0379 -
1.2964 12700 0.0284 -
1.3016 12750 0.0307 -
1.3067 12800 0.0191 -
1.3118 12850 0.031 -
1.3169 12900 0.031 -
1.3220 12950 0.0229 -
1.3271 13000 0.041 0.0346
1.3322 13050 0.0478 -
1.3373 13100 0.0299 -
1.3424 13150 0.0257 -
1.3475 13200 0.0348 -
1.3526 13250 0.0188 -
1.3577 13300 0.0109 -
1.3628 13350 0.0211 -
1.3679 13400 0.0271 -
1.3730 13450 0.0329 -
1.3781 13500 0.0232 -
1.3832 13550 0.0344 -
1.3883 13600 0.0226 -
1.3934 13650 0.0217 -
1.3985 13700 0.0265 -
1.4036 13750 0.0319 -
1.4087 13800 0.032 -
1.4138 13850 0.0278 -
1.4189 13900 0.0199 -
1.4241 13950 0.0331 -
1.4292 14000 0.0277 0.0331
1.4343 14050 0.0348 -
1.4394 14100 0.0321 -
1.4445 14150 0.0286 -
1.4496 14200 0.032 -
1.4547 14250 0.0267 -
1.4598 14300 0.0402 -
1.4649 14350 0.0284 -
1.4700 14400 0.0312 -
1.4751 14450 0.0251 -
1.4802 14500 0.0179 -
1.4853 14550 0.0341 -
1.4904 14600 0.0136 -
1.4955 14650 0.0411 -
1.5006 14700 0.0226 -
1.5057 14750 0.0292 -
1.5108 14800 0.0381 -
1.5159 14850 0.0277 -
1.5210 14900 0.0225 -
1.5261 14950 0.0209 -
1.5312 15000 0.0197 0.0333
1.5363 15050 0.0425 -
1.5414 15100 0.0405 -
1.5465 15150 0.0191 -
1.5517 15200 0.0388 -
1.5568 15250 0.0253 -
1.5619 15300 0.0231 -
1.5670 15350 0.0226 -
1.5721 15400 0.0331 -
1.5772 15450 0.0145 -
1.5823 15500 0.0365 -
1.5874 15550 0.0349 -
1.5925 15600 0.0186 -
1.5976 15650 0.02 -
1.6027 15700 0.0178 -
1.6078 15750 0.0366 -
1.6129 15800 0.0317 -
1.6180 15850 0.0262 -
1.6231 15900 0.0167 -
1.6282 15950 0.0127 -
1.6333 16000 0.0272 0.0320
1.6384 16050 0.0213 -
1.6435 16100 0.0267 -
1.6486 16150 0.0299 -
1.6537 16200 0.0312 -
1.6588 16250 0.037 -
1.6639 16300 0.0243 -
1.6690 16350 0.0223 -
1.6742 16400 0.0195 -
1.6793 16450 0.0258 -
1.6844 16500 0.0218 -
1.6895 16550 0.0206 -
1.6946 16600 0.0193 -
1.6997 16650 0.0318 -
1.7048 16700 0.0288 -
1.7099 16750 0.0281 -
1.7150 16800 0.026 -
1.7201 16850 0.0257 -
1.7252 16900 0.0291 -
1.7303 16950 0.0303 -
1.7354 17000 0.0234 0.0321
1.7405 17050 0.0228 -
1.7456 17100 0.0285 -
1.7507 17150 0.0348 -
1.7558 17200 0.0213 -
1.7609 17250 0.0362 -
1.7660 17300 0.0374 -
1.7711 17350 0.0339 -
1.7762 17400 0.0356 -
1.7813 17450 0.0325 -
1.7864 17500 0.0266 -
1.7915 17550 0.0295 -
1.7967 17600 0.0214 -
1.8018 17650 0.0383 -
1.8069 17700 0.0292 -
1.8120 17750 0.0135 -
1.8171 17800 0.0379 -
1.8222 17850 0.0262 -
1.8273 17900 0.0268 -
1.8324 17950 0.019 -
1.8375 18000 0.0201 0.0306
1.8426 18050 0.024 -
1.8477 18100 0.0174 -
1.8528 18150 0.0227 -
1.8579 18200 0.016 -
1.8630 18250 0.0248 -
1.8681 18300 0.023 -
1.8732 18350 0.0265 -
1.8783 18400 0.0286 -
1.8834 18450 0.0154 -
1.8885 18500 0.0358 -
1.8936 18550 0.011 -
1.8987 18600 0.0168 -
1.9038 18650 0.0213 -
1.9089 18700 0.0217 -
1.9140 18750 0.0177 -
1.9192 18800 0.0393 -
1.9243 18850 0.0176 -
1.9294 18900 0.0279 -
1.9345 18950 0.0242 -
1.9396 19000 0.0415 0.0296
1.9447 19050 0.0281 -
1.9498 19100 0.0423 -
1.9549 19150 0.0188 -
1.9600 19200 0.0417 -
1.9651 19250 0.0161 -
1.9702 19300 0.0277 -
1.9753 19350 0.014 -
1.9804 19400 0.0231 -
1.9855 19450 0.0344 -
1.9906 19500 0.0129 -
1.9957 19550 0.0241 -

Framework Versions

  • Python: 3.10.14
  • Sentence Transformers: 3.3.0
  • Transformers: 4.46.2
  • PyTorch: 2.4.0+cu121
  • Accelerate: 1.1.1
  • Datasets: 3.1.0
  • Tokenizers: 0.20.3

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",
}
Downloads last month
2
Safetensors
Model size
434M params
Tensor type
F32
·
Inference Examples
Sentence Similarity
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.