README.md

metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - dense
  - generated_from_trainer
  - dataset_size:12753278
  - loss:MarginMSELoss
base_model: PaDaS-Lab/xlm-roberta-base-msmarco
widget:
  - source_sentence: >-
      De combien de cm dois-je faire dépasser mon appui de fenêtre par rapport à
      mon mur ?
    sentences:
      - >-
        La profondeur idéale dépend de la configuration de l’endroit où vous
        l’installez et de vos goûts personnels. Si le radiateur doit être
        recouvert, un appui de fenêtre plus profond est nécessaire, tandis qu'un
        rebord plus étroit convient parfaitement à un style minimaliste. 5
        centimètres constituent une excellente base et sont suffisants dans la
        plupart des cas. Si la profondeur est plus importante, assurez-vous
        qu'au moins la moitié du rebord de la fenêtre repose sur le mur afin
        d'offrir une stabilité suffisante et de résister aux contraintes de la
        vie quotidienne. En savoir plus sur les tailles disponibles.
      - >-
        Pour les fenêtres sans jingot, il est obligatoire d'installer une bible
        qui couvre toute l'épaisseur des murs. Un écart de 10 mm rectiligne et
        15 mm rectiligne doit être respecté entre le nez de la bavette et le
        mur. .
      - >-
        För att registrera ett spelkonto och betta online måste du vara över 18
        år i Sverige.
  - source_sentence: Hvor kendt er Paulo Coelho som forfatter?
    sentences:
      - >-
        Paulo Coelho er en af verdens mest læste forfattere. Han er mest kendt
        for sin bog “Alkymisten”, som er blevet solgt i over 65 millioner
        eksemplarer og oversat til 71 forskellige sprog. Hans internationale
        succes har gjort ham til en af vor tids mest indflydelsesrige personer.
      - >-
        Sie spielen auf einem Raster von 5 Walzen und drei Reihen mit 25
        Gewinnlinien. Beim Blood Suckers Spiel gewinnen Sie von links nach
        rechts.
      - >-
        Nogle af forfatterens andre populære bøger inkluderer “Broen til det
        andet menneske” fra 2016, “Robuste børn” fra 2017 og “Opdragelse til
        livsmod og bæredygtighed” fra 2020. Disse bøger udforsker emner som
        medmenneskelighed, børneopdragelse og hvordan man kan skabe robuste og
        velfungerende børn.
  - source_sentence: >-
      Condominio minimo di due unità di proprietà della stessa persona. Una
      unità è posseduta dal proprietario e una dal locatario in virtù di giusto
      contratto di locazione. Possono intervenire sulle parti comuni
      naturalmente ognuno per le parti possedute?
    sentences:
      - Взносы и выплаты средств осуществляются в течение 1 рабочего дня.
      - "04/09/2020\nIl quesito non si specifica il tipo di agevolazione che si intende utilizzare.\nSe ipotizziamo l'utilizzo del Superbonus 110% del DL Rilancio convertito in L.77/2020 la risposta è NO.\nInfatti, in merito all'ambito soggettivo di applicazione dell'art.119 del DL Rilancio, tenuto conto della locuzione utilizzata dal legislatore riferita espressamente ai «condomini», ai fini dell’applicazione dell’agevolazione l’edificio oggetto degli interventi deve essere costituito in condominio secondo la disciplina civilistica prevista (secondo una consolidata giurisprudenza, la nascita del condominio si determina automaticamente, senza che sia necessaria alcuna deliberazione, nel momento in cui più soggetti costruiscono su un suolo comune ovvero quando l’unico proprietario di un edificio ne cede a terzi, piani o porzioni di piano, in proprietà esclusiva, realizzando l’oggettiva condizione del frazionamento, come chiarito dalla prassi in materia).\nIl locatario non può essere qualificato come condomino non essendo proprietario (di conseguenza neanche delle parti comuni) e pertanto, nel caso in specie, non costituendosi il Condominio, si ricade nel caso negativo dell'unico proprietario richiamato dalla Circolare AdE n.24/E del 08/08/2020 e cioè che il Superbonus non si applica agli interventi realizzati sulle parti comuni a due o più unità immobiliari distintamente accatastate di un edificio interamente posseduto da un unico \_proprietario o in comproprietà fra più soggetti."
      - "Su Marzo 18, 2013 la Legge sul diritto di Famiglia ha sostituito l'attuale Famiglia Relations Act, causando notevoli modifiche al diritto di famiglia in British Columbia. Uno dei maggiori cambiamenti è che di common law le coppie saranno governati dalla stessa divisione di beni, come le coppie sposate dopo la rottura della loro relazione. Diritto comune, le coppie, per le finalità di divisione di beni in il Diritto di Famiglia, Legge, comprende due persone che hanno vissuto in un matrimonio come rapporto per almeno due anni. La presunzione è che dopo un rapporto di ripartizione per ogni socio avrà diritto a uno la metà del â\x80di famiglia propertyâ\x80\x99 e aumenta di valore di â\x80escluso propertyâ\x80\x99. Se si desidera impostare un diverso accordo per la divisione dei beni, quindi un contratto di convivenza è il modo per farlo. Un contratto di convivenza può anche includere i dettagli di come ciascuna delle parti contribuirà finanziariamente durante il rapporto e i termini di supporto sponsale, se presente, il rapporto di composizione. Ogni contratto di convivenza deve essere specifico per la tua situazione, quindi, è essenziale rivolgersi ad un avvocato per un consiglio prima di stesura e stipula di eventuali accordi."
  - source_sentence: क्या लॉजिकल रीजनिंग यूजीसी नेट परीक्षा का हिस्सा है?
    sentences:
      - >-
        यूजीसी नेट की परीक्षा साल में दो बार आयोजित की जाती है। प्रथम परीक्षा
        जून में और द्वितीय परीक्षा दिसंबर महीने में नेशनल टेस्टिंग एजेंसी द्वारा
        आयोजित की जाती है।
      - >-
        जबकि मौलिक विश्लेषण किसी परिसंपत्ति के आंतरिक मूल्य का आकलन करता है,
        तकनीकी विश्लेषण पूरी तरह से परिसंपत्ति की कीमत के सांख्यिकीय रुझानों पर
        केंद्रित होता है। मौलिक विश्लेषण कंपनी की कमाई, आर्थिक संकेतक और प्रबंधन
        जैसे कारकों पर विचार करता है, जबकि तकनीकी विश्लेषण चार्ट और सांख्यिकीय
        रुझानों का उपयोग करता है।
      - हां, लॉजिकल रीजनिंग यूजीसी नेट परीक्षा का हिस्सा है।
  - source_sentence: >-
      Impact of the physico-chemical properties of fen peat on the metal
      accumulation patterns in mires of Latvia
    sentences:
      - >-
        Sì, EpiCURA ha 101 posizioni aperte. Prima di candidarsi presso EpiCURA,
        è opportuno fare ricerche sull'azienda e leggere le recensioni dei
        dipendenti che vi lavorano.
      - >-
        Wyckoff Method identifies accumulation and distribution phases by
        analyzing price and volume patterns. The goal is to locate support and
        resistance levels and patterns such as Springs and Upthrusts. Volume
        trends and moving averages provide confirmation.
      - "Santrauka\nAbstract The article presents a study of the physico-chemical properties of fen peat and their influence on the metal accumulation patterns in three Latvian fens: Svētupes Mire, Elku Mire and Vīķu Mire. Full peat profiles were obtained at all study sites and analysed with a multi-proxy approach. The content of metals in fen peat was determined using the atomic absorption spectroscopy (AAS) and normalised to the concentration of Ti in the studied peat profiles. Both the character of deposits and agricultural land use in the mire catchment areas were taken into account and the possible natural and anthropogenic metal supply sources were evaluated. The content of metals in the studied fen peat significantly varied due to the heterogeneity of fen environment; however, noticeable similarities were also traced throughout all study sites. The results indicate an increased amount of transition metals and Pb in the upper peat layer. This can be explained by a direct impact from anthropogenic sources (agricultural land use, pollution, etc.). Metal binding in fen peat profiles is directly related to the alkali and alkaline earth metal content in peat, as Ca, Mg, Na and K ions are replaced by more tightly bound metal ions. In raised bogs, in turn, metal binding is associated with the acidic functional groups common to peat.\nDoi\_10.5200/baltica.2016.29.03Raktažodžiai\_fen peat, metals, peat physico-chemical properties\nPilnas tekstas"
pipeline_tag: sentence-similarity
library_name: sentence-transformers

SentenceTransformer based on PaDaS-Lab/xlm-roberta-base-msmarco

This is a sentence-transformers model finetuned from PaDaS-Lab/xlm-roberta-base-msmarco. 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: PaDaS-Lab/xlm-roberta-base-msmarco
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False, 'architecture': 'XLMRobertaModel'})
  (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})
)

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("sentence_transformers_model_id")
# Run inference
sentences = [
    'Impact of the physico-chemical properties of fen peat on the metal accumulation patterns in mires of Latvia',
    'Santrauka\nAbstract The article presents a study of the physico-chemical properties of fen peat and their influence on the metal accumulation patterns in three Latvian fens: Svētupes Mire, Elku Mire and Vīķu Mire. Full peat profiles were obtained at all study sites and analysed with a multi-proxy approach. The content of metals in fen peat was determined using the atomic absorption spectroscopy (AAS) and normalised to the concentration of Ti in the studied peat profiles. Both the character of deposits and agricultural land use in the mire catchment areas were taken into account and the possible natural and anthropogenic metal supply sources were evaluated. The content of metals in the studied fen peat significantly varied due to the heterogeneity of fen environment; however, noticeable similarities were also traced throughout all study sites. The results indicate an increased amount of transition metals and Pb in the upper peat layer. This can be explained by a direct impact from anthropogenic sources (agricultural land use, pollution, etc.). Metal binding in fen peat profiles is directly related to the alkali and alkaline earth metal content in peat, as Ca, Mg, Na and K ions are replaced by more tightly bound metal ions. In raised bogs, in turn, metal binding is associated with the acidic functional groups common to peat.\nDoi\xa010.5200/baltica.2016.29.03Raktažodžiai\xa0fen peat, metals, peat physico-chemical properties\nPilnas tekstas',
    'Wyckoff Method identifies accumulation and distribution phases by analyzing price and volume patterns. The goal is to locate support and resistance levels and patterns such as Springs and Upthrusts. Volume trends and moving averages provide confirmation.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.9981, 0.9960],
#         [0.9981, 1.0000, 0.9978],
#         [0.9960, 0.9978, 1.0000]])

Training Details

Training Dataset

Unnamed Dataset

• Size: 12,753,278 training samples
• Columns: sentence_0, sentence_1, sentence_2, and label
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1	sentence_2	label
  type	string	string	string	float
  details	min: 6 tokens mean: 15.49 tokens max: 119 tokens	min: 9 tokens mean: 74.63 tokens max: 512 tokens	min: 11 tokens mean: 107.12 tokens max: 512 tokens	min: -0.48 mean: 0.16 max: 0.82

• Samples:

  sentence_0	sentence_1	sentence_2	label
  Как найти актуальное зеркало Betwinner?	Букмекер старается обеспечить доступ к сайту, поэтому ссылки на зеркала обновляются ежедневно. Чтобы быть в курсе всех новостей, рекомендуется подписаться на почтовую рассылку и соцсети.	Актуальные рабочие зеркала 1win можно найти на официальных страницах букмекерской конторы в социальных сетях или путем обращения в службу поддержки.	-0.253662109375
  Jakie są minimalne zakłady w bakaracie?	Minimalny zakład w bakaracie zależy od konkretnej gry, w którą grasz. Mini Baccarat ma zwykle niskie limity zakładów, co czyni go atrakcyjnym dla nowych graczy. Istnieją też wersje gry w bakarata dla high-rollerów, które nakładają wyższy minimalny zakład.	W czasie pisania tego tekstu, Cloudbet miał minimalny zakład w wysokości 0,01 BTC. Cloudbet ma jedne z najwyższych dostępnych maksymalnych zakładów, ale zazwyczaj różnią się one w zależności od płynności na dane zdarzenie. Ogólnie rzecz biorąc, im bliższe zdarzeniu zakłady, tym większa płynność i tym wyższe stawki można postawić.	0.3369140625
  Come scegliere il massimale assicurazione professionale medici?	Per il momento, non sono ancora entrate in vigore sul massimale minimo per le polizze rc professionale medici. Teniamo conto però di una cosa: se si lavora (e si è lavorato nei dieci anni precedenti) esclusivamente come dipendenti o specializzandi presso l’SSN, dobbiamo sapere che la rivalsa massima dell’SSN sarà plafonata al triplo del reddito annuo lordo del medico. Se invece si lavora in libera professione, non c’è alcun limite. Consigliamo comunque di scegliere massimali non inferiori al milione di euro.	Potete stipulare una nuova assicurazione di base entro il 31 dicembre, che inizierà a decorrere dal 1 gennaio dell’anno successivo. A tal fine, però, dovete aver disdetto la precedente assicurazione di base entro i termini previsti. In linea di massima è possibile disdire l’assicurazione di base inviando una lettera di disdetta firmata per posta, via e-mail o via fax. Alla KPT potete comunicare la disdetta anche nel portale clienti KPTnet. Attenzione: Per la disdetta non fa fede il timbro postale, bensì la data in cui l’assicuratore riceve la disdetta. Il termine di disdetta è rispettato se l’assicuratore riceve la disdetta l’ultimo giorno lavorativo del termine legale durante i normali orari di ufficio. Un invio raccomandato recapitato in una casella postale può essere considerato notificato solo nel momento in cui è ritirato alla Posta. Poiché non si possono escludere ritardi, vi consigliamo di inviare la disdetta per posta, mediante lettera raccomandata, entro la metà di novembre. C...	0.09228515625

• Loss: MarginMSELoss

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• num_train_epochs: 1
• fp16: True
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: no
• prediction_loss_only: True
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 5e-05
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1
• num_train_epochs: 1
• 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
• bf16: False
• fp16: True
• 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}
• parallelism_config: 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
• project: huggingface
• trackio_space_id: trackio
• 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
• hub_revision: None
• 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
• include_tokens_per_second: False
• include_num_input_tokens_seen: no
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• eval_on_start: False
• use_liger_kernel: False
• liger_kernel_config: None
• eval_use_gather_object: False
• average_tokens_across_devices: True
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: round_robin
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Click to expand

Epoch	Step	Training Loss
0.0025	500	4.076
0.0050	1000	0.0295
0.0075	1500	0.0259
0.0100	2000	0.0242
0.0125	2500	0.0226
0.0151	3000	0.022
0.0176	3500	0.0221
0.0201	4000	0.0211
0.0226	4500	0.0208
0.0251	5000	0.0204
0.0276	5500	0.0202
0.0301	6000	0.0197
0.0326	6500	0.0196
0.0351	7000	0.0194
0.0376	7500	0.0193
0.0401	8000	0.0191
0.0427	8500	0.0195
0.0452	9000	0.02
0.0477	9500	0.0189
0.0502	10000	0.0184
0.0527	10500	0.0185
0.0552	11000	0.0183
0.0577	11500	0.0191
0.0602	12000	0.018
0.0627	12500	0.0177
0.0652	13000	0.0177
0.0677	13500	0.0175
0.0703	14000	0.0174
0.0728	14500	0.0172
0.0753	15000	0.0175
0.0778	15500	0.0171
0.0803	16000	0.0174
0.0828	16500	0.0175
0.0853	17000	0.0169
0.0878	17500	0.0168
0.0903	18000	0.0167
0.0928	18500	0.0171
0.0953	19000	0.0169
0.0979	19500	0.0167
0.1004	20000	0.0163
0.1029	20500	0.0164
0.1054	21000	0.0168
0.1079	21500	0.0163
0.1104	22000	0.0167
0.1129	22500	0.0162
0.1154	23000	0.0163
0.1179	23500	0.0159
0.1204	24000	0.0163
0.1229	24500	0.0159
0.1255	25000	0.0161
0.1280	25500	0.0161
0.1305	26000	0.0159
0.1330	26500	0.0159
0.1355	27000	0.0159
0.1380	27500	0.0158
0.1405	28000	0.0157
0.1430	28500	0.0157
0.1455	29000	0.0156
0.1480	29500	0.0172
0.1505	30000	0.0155
0.1531	30500	0.0153
0.1556	31000	0.0152
0.1581	31500	0.0154
0.1606	32000	0.0153
0.1631	32500	0.0153
0.1656	33000	0.0153
0.1681	33500	0.0153
0.1706	34000	0.0151
0.1731	34500	0.015
0.1756	35000	0.0148
0.1782	35500	0.015
0.1807	36000	0.0148
0.1832	36500	0.0149
0.1857	37000	0.0147
0.1882	37500	0.0145
0.1907	38000	0.0145
0.1932	38500	0.0147
0.1957	39000	0.0149
0.1982	39500	0.0145
0.2007	40000	0.0145
0.2032	40500	0.0147
0.2058	41000	0.0147
0.2083	41500	0.0147
0.2108	42000	0.0145
0.2133	42500	0.0144
0.2158	43000	0.0147
0.2183	43500	0.0145
0.2208	44000	0.0147
0.2233	44500	0.0142
0.2258	45000	0.0145
0.2283	45500	0.0141
0.2308	46000	0.0143
0.2334	46500	0.0143
0.2359	47000	0.0141
0.2384	47500	0.0145
0.2409	48000	0.0142
0.2434	48500	0.0141
0.2459	49000	0.0142
0.2484	49500	0.0139
0.2509	50000	0.0141
0.2534	50500	0.0139
0.2559	51000	0.014
0.2584	51500	0.0139
0.2610	52000	0.014
0.2635	52500	0.0142
0.2660	53000	0.014
0.2685	53500	0.0138
0.2710	54000	0.0136
0.2735	54500	0.0138
0.2760	55000	0.0138
0.2785	55500	0.0137
0.2810	56000	0.0136
0.2835	56500	0.0138
0.2860	57000	0.0135
0.2886	57500	0.0135
0.2911	58000	0.0137
0.2936	58500	0.0136
0.2961	59000	0.0135
0.2986	59500	0.0143
0.3011	60000	0.0134
0.3036	60500	0.0135
0.3061	61000	0.0136
0.3086	61500	0.0134
0.3111	62000	0.0134
0.3136	62500	0.0132
0.3162	63000	0.0133
0.3187	63500	0.0133
0.3212	64000	0.0135
0.3237	64500	0.0133
0.3262	65000	0.0133
0.3287	65500	0.0134
0.3312	66000	0.0133
0.3337	66500	0.0132
0.3362	67000	0.0133
0.3387	67500	0.0133
0.3412	68000	0.0132
0.3438	68500	0.0131
0.3463	69000	0.0132
0.3488	69500	0.0131
0.3513	70000	0.013
0.3538	70500	0.0129
0.3563	71000	0.0127
0.3588	71500	0.0131
0.3613	72000	0.0129
0.3638	72500	0.0128
0.3663	73000	0.0129
0.3688	73500	0.0128
0.3714	74000	0.0128
0.3739	74500	0.0131
0.3764	75000	0.013
0.3789	75500	0.0127
0.3814	76000	0.0128
0.3839	76500	0.0127
0.3864	77000	0.0128
0.3889	77500	0.0129
0.3914	78000	0.0128
0.3939	78500	0.0127
0.3964	79000	0.0128
0.3990	79500	0.0126
0.4015	80000	0.0127
0.4040	80500	0.0126
0.4065	81000	0.0124
0.4090	81500	0.0126
0.4115	82000	0.0124
0.4140	82500	0.0124
0.4165	83000	0.0127
0.4190	83500	0.0123
0.4215	84000	0.0124
0.4240	84500	0.0125
0.4266	85000	0.0124
0.4291	85500	0.0124
0.4316	86000	0.0124
0.4341	86500	0.0124
0.4366	87000	0.0128
0.4391	87500	0.0124
0.4416	88000	0.0123
0.4441	88500	0.0123
0.4466	89000	0.0125
0.4491	89500	0.0125
0.4516	90000	0.0123
0.4542	90500	0.0124
0.4567	91000	0.0122
0.4592	91500	0.0122
0.4617	92000	0.0124
0.4642	92500	0.012
0.4667	93000	0.0122
0.4692	93500	0.0121
0.4717	94000	0.0121
0.4742	94500	0.0121
0.4767	95000	0.0123
0.4792	95500	0.0121
0.4818	96000	0.0121
0.4843	96500	0.0127
0.4868	97000	0.012
0.4893	97500	0.0122
0.4918	98000	0.012
0.4943	98500	0.0119
0.4968	99000	0.012
0.4993	99500	0.0121
0.5018	100000	0.012
0.5043	100500	0.0119
0.5069	101000	0.0121
0.5094	101500	0.0123
0.5119	102000	0.0117
0.5144	102500	0.0121
0.5169	103000	0.0118
0.5194	103500	0.0118
0.5219	104000	0.0118
0.5244	104500	0.0119
0.5269	105000	0.012
0.5294	105500	0.0117
0.5319	106000	0.0118
0.5345	106500	0.0118
0.5370	107000	0.0118
0.5395	107500	0.0119
0.5420	108000	0.0116
0.5445	108500	0.012
0.5470	109000	0.0116
0.5495	109500	0.0116
0.5520	110000	0.0116
0.5545	110500	0.0117
0.5570	111000	0.0117
0.5595	111500	0.0117
0.5621	112000	0.0116
0.5646	112500	0.0116
0.5671	113000	0.0116
0.5696	113500	0.0116
0.5721	114000	0.0116
0.5746	114500	0.012
0.5771	115000	0.0119
0.5796	115500	0.0115
0.5821	116000	0.0116
0.5846	116500	0.0115
0.5871	117000	0.0116
0.5897	117500	0.0116
0.5922	118000	0.0115
0.5947	118500	0.0116
0.5972	119000	0.0115
0.5997	119500	0.0116
0.6022	120000	0.0114
0.6047	120500	0.0115
0.6072	121000	0.0115
0.6097	121500	0.0114
0.6122	122000	0.0115
0.6147	122500	0.0114
0.6173	123000	0.0113
0.6198	123500	0.0112
0.6223	124000	0.0114
0.6248	124500	0.0113
0.6273	125000	0.0112
0.6298	125500	0.0115
0.6323	126000	0.0112
0.6348	126500	0.0112
0.6373	127000	0.0113
0.6398	127500	0.0113
0.6423	128000	0.0113
0.6449	128500	0.0112
0.6474	129000	0.0111
0.6499	129500	0.0114
0.6524	130000	0.0111
0.6549	130500	0.0111
0.6574	131000	0.0112
0.6599	131500	0.0111
0.6624	132000	0.0113
0.6649	132500	0.0112
0.6674	133000	0.0112
0.6699	133500	0.0111
0.6725	134000	0.0111
0.6750	134500	0.0111
0.6775	135000	0.011
0.6800	135500	0.0113
0.6825	136000	0.011
0.6850	136500	0.011
0.6875	137000	0.0111
0.6900	137500	0.0111
0.6925	138000	0.0112
0.6950	138500	0.0112
0.6975	139000	0.0109
0.7001	139500	0.0112
0.7026	140000	0.011
0.7051	140500	0.011
0.7076	141000	0.0108
0.7101	141500	0.0109
0.7126	142000	0.0108
0.7151	142500	0.0109
0.7176	143000	0.0109
0.7201	143500	0.0109
0.7226	144000	0.0112
0.7251	144500	0.011
0.7277	145000	0.0108
0.7302	145500	0.0109
0.7327	146000	0.0111
0.7352	146500	0.0109
0.7377	147000	0.0109
0.7402	147500	0.0108
0.7427	148000	0.011
0.7452	148500	0.0108
0.7477	149000	0.0109
0.7502	149500	0.0107
0.7527	150000	0.0108
0.7553	150500	0.011
0.7578	151000	0.0107
0.7603	151500	0.0108
0.7628	152000	0.0107
0.7653	152500	0.0108
0.7678	153000	0.011
0.7703	153500	0.0108
0.7728	154000	0.0108
0.7753	154500	0.0108
0.7778	155000	0.0106
0.7803	155500	0.0107
0.7829	156000	0.0107
0.7854	156500	0.0107
0.7879	157000	0.0107
0.7904	157500	0.0106
0.7929	158000	0.0107
0.7954	158500	0.0107
0.7979	159000	0.0107
0.8004	159500	0.0107
0.8029	160000	0.0105
0.8054	160500	0.0106
0.8079	161000	0.0106
0.8105	161500	0.0108
0.8130	162000	0.0107
0.8155	162500	0.0106
0.8180	163000	0.0107
0.8205	163500	0.0106
0.8230	164000	0.0107
0.8255	164500	0.0105
0.8280	165000	0.0107
0.8305	165500	0.0107
0.8330	166000	0.0107
0.8355	166500	0.0105
0.8381	167000	0.0106
0.8406	167500	0.0105
0.8431	168000	0.0106
0.8456	168500	0.0105
0.8481	169000	0.0105
0.8506	169500	0.0106
0.8531	170000	0.0105
0.8556	170500	0.0105
0.8581	171000	0.0105
0.8606	171500	0.0106
0.8632	172000	0.0106
0.8657	172500	0.0104
0.8682	173000	0.0106
0.8707	173500	0.0105
0.8732	174000	0.0105
0.8757	174500	0.0104
0.8782	175000	0.0104
0.8807	175500	0.0106
0.8832	176000	0.0106
0.8857	176500	0.0103
0.8882	177000	0.0104
0.8908	177500	0.0104
0.8933	178000	0.0105
0.8958	178500	0.0105
0.8983	179000	0.0102
0.9008	179500	0.0105
0.9033	180000	0.0104
0.9058	180500	0.0104
0.9083	181000	0.0103
0.9108	181500	0.0104
0.9133	182000	0.0104
0.9158	182500	0.0103
0.9184	183000	0.0104
0.9209	183500	0.0104
0.9234	184000	0.0103
0.9259	184500	0.0105
0.9284	185000	0.0103
0.9309	185500	0.0103
0.9334	186000	0.0106
0.9359	186500	0.0103
0.9384	187000	0.0108
0.9409	187500	0.0103
0.9434	188000	0.0103
0.9460	188500	0.0103
0.9485	189000	0.0105
0.9510	189500	0.0104
0.9535	190000	0.0102
0.9560	190500	0.0102
0.9585	191000	0.0103
0.9610	191500	0.0101
0.9635	192000	0.0103
0.9660	192500	0.0105
0.9685	193000	0.0102
0.9710	193500	0.0102
0.9736	194000	0.0103
0.9761	194500	0.0102
0.9786	195000	0.0102
0.9811	195500	0.0102
0.9836	196000	0.0104
0.9861	196500	0.0103
0.9886	197000	0.0103
0.9911	197500	0.0102
0.9936	198000	0.0103
0.9961	198500	0.0101
0.9986	199000	0.0102

Framework Versions

• Python: 3.10.4
• Sentence Transformers: 5.2.0
• Transformers: 4.57.3
• PyTorch: 2.9.1+cu128
• Accelerate: 1.12.0
• Datasets: 2.21.0
• Tokenizers: 0.22.1

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",
}

MarginMSELoss

@misc{hofstätter2021improving,
    title={Improving Efficient Neural Ranking Models with Cross-Architecture Knowledge Distillation},
    author={Sebastian Hofstätter and Sophia Althammer and Michael Schröder and Mete Sertkan and Allan Hanbury},
    year={2021},
    eprint={2010.02666},
    archivePrefix={arXiv},
    primaryClass={cs.IR}
}