thebajajra/RexBERT-base-embed-pf-v0.4

SentenceTransformer based on thebajajra/RexBERT-base-embed-pf-v0.1

This is a sentence-transformers model finetuned from thebajajra/RexBERT-base-embed-pf-v0.1 on the nomic-embed-supervised-data 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: thebajajra/RexBERT-base-embed-pf-v0.1
• Maximum Sequence Length: 1024 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • nomic-embed-supervised-data

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': 1024, 'do_lower_case': False, 'architecture': 'ModernBertModel'})
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
)

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
queries = [
    "None of this proves that Loral wasn\u0027t disloyal or criminally negligent in its dealings with China.",
]
documents = [
    'There is nothing that proves that Loral was negligent in dealing with China. ',
    'Trimipramine is a tricyclic antidepressant. Trimipramine affects chemicals in the brain that may become unbalanced. Trimipramine is used to treat symptoms of depression.Trimipramine may also be used for purposes not listed in this medication guide. You should not take trimipramine if you have recently had a heart attack. Do not use trimipramine if you have used an MAO inhibitor in the past 14 days.A dangerous drug interaction could occur.ou should not use trimipramine if you are allergic to it, or if you have: 1  if you have recently had a heart attack; or. 2  if you are allergic to antidepressants such as amitriptyline, amoxapine, clomipramine, desipramine, doxepin, imipramine, nortriptyline, or protriptyline.',
    'Lorraine Dunn Lorraine Dunn (12 September 1942 – 16 October 2003) was a Panamanian sprinter. She competed in the 4 × 100 metres relay at the 1960 Summer Olympics and the 1964 Summer Olympics. Dunn finished third in the 200 metres and finished fourth in the 80 metres hurdles at the 1963 Pan American Games.',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 768] [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)
# tensor([[ 0.9269, -0.0305,  0.0724]])

Training Details

Training Dataset

nomic-embed-supervised-data

• Dataset: nomic-embed-supervised-data at 13eef8a
• Size: 1,611,024 training samples
• Columns: query, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	query	positive	negative
  type	string	string	list
  details	min: 4 tokens mean: 37.16 tokens max: 1024 tokens	min: 5 tokens mean: 99.03 tokens max: 1024 tokens	min: 20 elements mean: 155.14 elements max: 209 elements

• Samples:

  query	positive	negative
  Japan's biggest daily newspaper, Asahi Shimbun, published an interview Sunday with Woody Allen, who, asked to sum up the 20 th century in a single word, replied, Disappointing.	Woody Allen described the 20th century as "disappointing'' in an interview with the Japaense newspaper, Asahi Shimbun.	['Billy Joel described the 20th century as "outlandish" in a recent interview with Japan's leading newspaper. ', "aside for the kids' education and--BOP--I have so much trouble with that the", 'An old man is jogging.', 'It was my favorite spot.', 'Two women are playing volleyball.', ...]
  Sedimentation is the term for when clumped solids sink to the bottom of the water.	Deposition occurs where the water motion slows.	['The law of superposition is best described by: in undisturbed layers of sedimentary rock, the lowest layers contain the older rocks.', "Review: I felt this film - throughout. I waas impressed with Russell Crowe's talent in developing his relationship with Lillie, such a typical Aussie blend of softly softly approach, a bit self depreciating and very persistent. Really loved the cinematography and direction. Pace was just right and the portrayals of nearly all characters was impressive.Gosh, didn't Russell's talent even in 1993 shine! .. and I have yet to see Gladiator. Question: is it a negative review?", 'Question: How many times did Chopin and Liszy perform together in public? (Answer: seven).', 'Fact 1: Birth may be followed by a period of parental care of the offspring. \nFact 2: Human birth is the rarest of all births.', 'acts. Id. at 963. The Lawson court stated that a "critical distinction” of Pylant from the facts before it in Lawson was that in Pylant, there was "not a factu...
  Volleyball involves techniques like jumping.	Volleyball Volleyball is a team sport in which two teams of six players are separated by a net . Each team tries to score points by grounding a ball on the other team 's court under organized rules . It has been a part of the official program of the Summer Olympic Games since 1964 . The complete rules are extensive . But simply , play proceeds as follows : a player on one of the teams begins a ` rally ' by serving the ball ( tossing or releasing it and then hitting it with a hand or arm ) , from behind the back boundary line of the court , over the net , and into the receiving team 's court . The receiving team must not let the ball be grounded within their court . The team may touch the ball up to 3 times but individual players may not touch the ball twice consecutively . Typically , the first two touches are used to set up for an attack , an attempt to direct the ball back over the net in such a way that the serving team is unable to prevent it from being grounded in their court . ...	['Volleyball jump serve The Volleyball jump Serve is a type of volleyball serve where the player increases the power and height of their serve by jumping into the hit . The Jump Serve itself was popularized by the brazilian national volleyball team in 1984 on the Olympics .', "Volleyball variations As volleyball is one of the world 's most popular team sports , second only to football ( soccer ) in the number of players , there are numerous variations of the basic rules . The rules have changed around the world since its creation in 1895 , as skills have developed , to make the game more suited for spectators , for learning or other special needs . Some variations have been included as a change in the international rules by Fédération Internationale de Volleyball ( FIVB ) , others have resulted in variations with specific names . The more notable variations include :", 'Volleyball (ball) A volleyball is a ball used to play indoor volleyball , beach volleyball , or other less common v...

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false
  }
  

Evaluation Dataset

nomic-embed-supervised-data

• Dataset: nomic-embed-supervised-data at 13eef8a
• Size: 84,795 evaluation samples
• Columns: query, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	query	positive	negative
  type	string	string	list
  details	min: 4 tokens mean: 34.78 tokens max: 1024 tokens	min: 5 tokens mean: 91.63 tokens max: 1024 tokens	min: 15 elements mean: 151.78 elements max: 209 elements

• Samples:

  query	positive	negative
  I think people are upset that they're "copying" a mechanic. Which is ridiculous. Every platformer copied from Mario. That doesn't mean they weren't good games. And this says nothing of the end result of the flow of the game.	The people who say spiderman is copying another game. Are the same retarded little children who think any game that has drivable vehicles in it, is a gta rip off. You can't take them serious.	['i unfollowed all of them earlier this year theres too many and all their tweets and instaposts made up damn near half my feed i said id just wait for the music and then iridesence came out and was pretty mid so ', 'There are five, FIVE fucking boxes of doughnuts by the printer in the office this morning and I am overtired and hungry and in the worst mood, and I’m sorry r/1200isplenty, tea is not a goddamn dessert substitute; I keep drinking it and not only am I not satiated, I have to pee every ten minutes, so I’m here guzzling tea and growling ”nothing tastes as good as skinny feels” under my breath and if I sound crazy, it’s because I am. \n\nEdit: down to one box. Just checked the calories for one doughnut—380. Three hundred. And eighty. Calories. For one doughnut. ', 'What a strange thing to call your cock.', "It doesn't. That's a very personal decision. You're not right or wrong on that. I was just replying because you were clear you don't want to raise the child as your own....
  duties of a medical assistant for resume	Include clinical skills in a health care resume. Those that apply to a medical assistant job may be: 1 Preparing patients, including taking of medical histories. 2 The ability to take vital signs. 3 Preparing medication and treatments. 4 Basic first aid, CPR and infection control knowledge. Assisting physicians with exams.	['Include clinical skills in a health care resume. Those that apply to a medical assistant job may be: 1 Preparing patients, including taking of medical histories. 2 The ability to take vital signs. 3 Preparing medication and treatments. Basic first aid, CPR and infection control knowledge.', 'Clinical Medical Assistant Resume. The position of a medical assistant is a key position in any hospital, clinic or a health care center. A medical assistant may perform a variety of duties, pertaining to administrative, front office or clinical duties. Here we will discuss the job profile of a clinical medical assistant. Also, there is a clinical medical assistant resume example provided below, for your further reference.', "Community Q&A. A medical assistant is a member of a health care team that undertakes administrative and clinical job duties. A health care resume must emphasize skills as well as education and job experience. Learn how to write a resume for a medical assistant job and yo...
  when did british rule end in south africa	History of South Africa British colonies: Cape Colony, Natal Colony, Transvaal Colony, and Orange River Colony. The country became a self-governing nation state within the British Empire, in 1934 following enactment of the Status of the Union Act. The dominion came to an end on 31 May 1961 as the consequence of a 1960 referendum, which legitimised the country becoming a sovereign state named Republic of South Africa. A republican constitution was adopted. From 1948–1994, South African politics were dominated by Afrikaner nationalism. Racial segregation and white minority rule known officially as apartheid, an Afrikaans word meaning "separateness”, came into existence in 1948 (under	["South Africa 1910. The Union was a dominion that included the former territories of the Cape, Transvaal and Natal colonies, as well as the Orange Free State republic. The Natives' Land Act of 1913 severely restricted the ownership of land by blacks; at that stage natives controlled only 7% of the country. The amount of land reserved for indigenous peoples was later marginally increased. In 1931, the union was fully sovereign from the United Kingdom with the passage of the Statute of Westminster, which abolished the last powers of the British Government on the country. In 1934, the South African Party and", 'History of South Africa conducted raids in Free State territories. Both sides adopted scorched-earth tactics, with large swathes of pasturage and cropland being destroyed. Faced with starvation, Moshoeshoe signed a peace treaty on 15 October 1858, though crucial boundary issues remained unresolved. War broke out again in 1865. After an unsuccessful appeal for aid from the British ...

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 256
• per_device_eval_batch_size: 128
• learning_rate: 2e-06
• num_train_epochs: 20
• warmup_ratio: 0.1
• bf16: True
• dataloader_num_workers: 10
• dataloader_prefetch_factor: 10

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 256
• per_device_eval_batch_size: 128
• 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-06
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 20
• max_steps: -1
• lr_scheduler_type: linear
• 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
• 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: True
• dataloader_num_workers: 10
• dataloader_prefetch_factor: 10
• 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: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Click to expand

Epoch	Step	Training Loss	Validation Loss
0.0127	10	1.2782	-
0.0254	20	1.2831	-
0.0382	30	1.2715	-
0.0509	40	1.2557	-
0.0636	50	1.3035	-
0.0763	60	1.2501	-
0.0891	70	1.2792	-
0.1018	80	1.2826	-
0.1145	90	1.2824	-
0.1272	100	1.3051	-
0.1399	110	1.2809	-
0.1527	120	1.2685	-
0.1654	130	1.2746	-
0.1781	140	1.2672	-
0.1908	150	1.2706	-
0.2036	160	1.2905	-
0.2163	170	1.268	-
0.2290	180	1.2641	-
0.2417	190	1.2823	-
0.2545	200	1.2985	-
0.2672	210	1.2695	-
0.2799	220	1.2772	-
0.2926	230	1.281	-
0.3053	240	1.2709	-
0.3181	250	1.2797	-
0.3308	260	1.2659	-
0.3435	270	1.2721	-
0.3562	280	1.2654	-
0.3690	290	1.2797	-
0.3817	300	1.2656	-
0.3944	310	1.2868	-
0.4071	320	1.2663	-
0.4198	330	1.2686	-
0.4326	340	1.2833	-
0.4453	350	1.3006	-
0.4580	360	1.2474	-
0.4707	370	1.2655	-
0.4835	380	1.2561	-
0.4962	390	1.2541	-
0.5	393	-	1.1388
0.5089	400	1.2582	-
0.5216	410	1.2614	-
0.5344	420	1.2473	-
0.5471	430	1.245	-
0.5598	440	1.2457	-
0.5725	450	1.2705	-
0.5852	460	1.2537	-
0.5980	470	1.2481	-
0.6107	480	1.245	-
0.6234	490	1.2499	-
0.6361	500	1.2403	-
0.6489	510	1.2334	-
0.6616	520	1.2484	-
0.6743	530	1.2255	-
0.6870	540	1.2394	-
0.6997	550	1.2209	-
0.7125	560	1.2383	-
0.7252	570	1.2522	-
0.7379	580	1.2261	-
0.7506	590	1.2171	-
0.7634	600	1.2461	-
0.7761	610	1.2123	-
0.7888	620	1.2174	-
0.8015	630	1.2326	-
0.8142	640	1.227	-
0.8270	650	1.2265	-
0.8397	660	1.2305	-
0.8524	670	1.2132	-
0.8651	680	1.2076	-
0.8779	690	1.2094	-
0.8906	700	1.2163	-
0.9033	710	1.1944	-
0.9160	720	1.2035	-
0.9288	730	1.1791	-
0.9415	740	1.1877	-
0.9542	750	1.1858	-
0.9669	760	1.1756	-
0.9796	770	1.1509	-
0.9924	780	1.1491	-
1.0	786	-	1.0399
1.0051	790	1.1459	-
1.0178	800	1.1585	-
1.0305	810	1.158	-
1.0433	820	1.1474	-
1.0560	830	1.1399	-
1.0687	840	1.1487	-
1.0814	850	1.138	-
1.0941	860	1.1133	-
1.1069	870	1.1085	-
1.1196	880	1.1271	-
1.1323	890	1.1232	-
1.1450	900	1.109	-
1.1578	910	1.0807	-
1.1705	920	1.0983	-
1.1832	930	1.0837	-
1.1959	940	1.0649	-
1.2087	950	1.0596	-
1.2214	960	1.0741	-
1.2341	970	1.0736	-
1.2468	980	1.062	-
1.2595	990	1.0579	-
1.2723	1000	1.0194	-
1.2850	1010	1.0345	-
1.2977	1020	1.0193	-
1.3104	1030	1.0243	-
1.3232	1040	0.9892	-
1.3359	1050	0.9933	-
1.3486	1060	0.9802	-
1.3613	1070	0.9755	-
1.3740	1080	0.9636	-
1.3868	1090	0.9487	-
1.3995	1100	0.9552	-
1.4122	1110	0.9436	-
1.4249	1120	0.925	-
1.4377	1130	0.9186	-
1.4504	1140	0.8905	-
1.4631	1150	0.8737	-
1.4758	1160	0.8613	-
1.4885	1170	0.8602	-
1.5	1179	-	0.7025
1.5013	1180	0.8286	-
1.5140	1190	0.8159	-
1.5267	1200	0.8142	-
1.5394	1210	0.7728	-
1.5522	1220	0.7741	-
1.5649	1230	0.755	-
1.5776	1240	0.7797	-
1.5903	1250	0.7565	-
1.6031	1260	0.7162	-
1.6158	1270	0.7182	-
1.6285	1280	0.7005	-
1.6412	1290	0.6993	-
1.6539	1300	0.694	-
1.6667	1310	0.6761	-
1.6794	1320	0.6702	-
1.6921	1330	0.6657	-
1.7048	1340	0.655	-
1.7176	1350	0.6436	-
1.7303	1360	0.615	-
1.7430	1370	0.6324	-
1.7557	1380	0.6312	-
1.7684	1390	0.6078	-
1.7812	1400	0.6058	-
1.7939	1410	0.5995	-
1.8066	1420	0.6053	-
1.8193	1430	0.5753	-
1.8321	1440	0.5769	-
1.8448	1450	0.5719	-
1.8575	1460	0.5822	-
1.8702	1470	0.5666	-
1.8830	1480	0.5544	-
1.8957	1490	0.5284	-
1.9084	1500	0.5393	-
1.9211	1510	0.5213	-
1.9338	1520	0.5298	-
1.9466	1530	0.513	-
1.9593	1540	0.5211	-
1.9720	1550	0.514	-
1.9847	1560	0.5019	-
1.9975	1570	0.4971	-
2.0	1572	-	0.3834
2.0102	1580	0.4904	-
2.0229	1590	0.5015	-
2.0356	1600	0.4777	-
2.0483	1610	0.4928	-
2.0611	1620	0.473	-
2.0738	1630	0.4781	-
2.0865	1640	0.4715	-
2.0992	1650	0.462	-
2.1120	1660	0.4628	-
2.1247	1670	0.4703	-
2.1374	1680	0.4709	-
2.1501	1690	0.4631	-
2.1628	1700	0.4563	-
2.1756	1710	0.4638	-
2.1883	1720	0.4571	-
2.2010	1730	0.4551	-
2.2137	1740	0.4353	-
2.2265	1750	0.4472	-
2.2392	1760	0.4539	-
2.2519	1770	0.4508	-
2.2646	1780	0.4529	-
2.2774	1790	0.4507	-
2.2901	1800	0.4593	-
2.3028	1810	0.4611	-
2.3155	1820	0.4455	-
2.3282	1830	0.4302	-
2.3410	1840	0.4352	-
2.3537	1850	0.4395	-
2.3664	1860	0.4282	-
2.3791	1870	0.4409	-
2.3919	1880	0.4243	-
2.4046	1890	0.4256	-
2.4173	1900	0.4419	-
2.4300	1910	0.4262	-
2.4427	1920	0.4254	-
2.4555	1930	0.4117	-
2.4682	1940	0.4294	-
2.4809	1950	0.4227	-
2.4936	1960	0.4192	-
2.5	1965	-	0.3127
2.5064	1970	0.4212	-
2.5191	1980	0.4288	-
2.5318	1990	0.4076	-
2.5445	2000	0.4158	-
2.5573	2010	0.4186	-
2.5700	2020	0.4213	-
2.5827	2030	0.4144	-
2.5954	2040	0.4343	-
2.6081	2050	0.4193	-
2.6209	2060	0.4149	-
2.6336	2070	0.4181	-
2.6463	2080	0.411	-
2.6590	2090	0.4142	-
2.6718	2100	0.4232	-
2.6845	2110	0.4083	-
2.6972	2120	0.4192	-
2.7099	2130	0.4233	-
2.7226	2140	0.4196	-
2.7354	2150	0.4265	-
2.7481	2160	0.4176	-
2.7608	2170	0.4165	-
2.7735	2180	0.3897	-
2.7863	2190	0.4119	-
2.7990	2200	0.4181	-
2.8117	2210	0.4214	-
2.8244	2220	0.4122	-
2.8372	2230	0.407	-
2.8499	2240	0.4068	-
2.8626	2250	0.4125	-
2.8753	2260	0.4153	-
2.8880	2270	0.4119	-
2.9008	2280	0.4168	-
2.9135	2290	0.4219	-
2.9262	2300	0.4149	-
2.9389	2310	0.4244	-
2.9517	2320	0.4052	-
2.9644	2330	0.4073	-
2.9771	2340	0.4116	-
2.9898	2350	0.4148	-
3.0	2358	-	0.2973
3.0025	2360	0.4002	-
3.0153	2370	0.4056	-
3.0280	2380	0.4086	-
3.0407	2390	0.4131	-
3.0534	2400	0.4066	-
3.0662	2410	0.4062	-
3.0789	2420	0.4116	-
3.0916	2430	0.3935	-
3.1043	2440	0.405	-
3.1170	2450	0.3993	-
3.1298	2460	0.4062	-
3.1425	2470	0.4027	-
3.1552	2480	0.3876	-
3.1679	2490	0.4075	-
3.1807	2500	0.4015	-
3.1934	2510	0.4137	-
3.2061	2520	0.4074	-
3.2188	2530	0.4134	-
3.2316	2540	0.4139	-
3.2443	2550	0.4106	-
3.2570	2560	0.4103	-
3.2697	2570	0.3893	-
3.2824	2580	0.4114	-
3.2952	2590	0.407	-
3.3079	2600	0.4025	-
3.3206	2610	0.3997	-
3.3333	2620	0.4054	-
3.3461	2630	0.4075	-
3.3588	2640	0.3926	-
3.3715	2650	0.3924	-
3.3842	2660	0.4245	-
3.3969	2670	0.4087	-
3.4097	2680	0.402	-
3.4224	2690	0.4094	-
3.4351	2700	0.4044	-
3.4478	2710	0.4131	-
3.4606	2720	0.4098	-
3.4733	2730	0.3998	-
3.4860	2740	0.3948	-
3.4987	2750	0.4009	-
3.5	2751	-	0.2923
3.5115	2760	0.3911	-
3.5242	2770	0.3964	-
3.5369	2780	0.4002	-
3.5496	2790	0.4055	-
3.5623	2800	0.4008	-
3.5751	2810	0.3956	-
3.5878	2820	0.4164	-
3.6005	2830	0.3971	-
3.6132	2840	0.3946	-
3.6260	2850	0.3974	-
3.6387	2860	0.3901	-
3.6514	2870	0.3997	-
3.6641	2880	0.4094	-
3.6768	2890	0.4031	-
3.6896	2900	0.4042	-
3.7023	2910	0.3977	-
3.7150	2920	0.3974	-
3.7277	2930	0.4059	-
3.7405	2940	0.3967	-
3.7532	2950	0.3981	-
3.7659	2960	0.3951	-
3.7786	2970	0.4044	-
3.7913	2980	0.3964	-
3.8041	2990	0.3925	-
3.8168	3000	0.3988	-
3.8295	3010	0.3913	-
3.8422	3020	0.4037	-
3.8550	3030	0.4008	-
3.8677	3040	0.3984	-
3.8804	3050	0.395	-
3.8931	3060	0.4005	-
3.9059	3070	0.3934	-
3.9186	3080	0.3858	-
3.9313	3090	0.3984	-
3.9440	3100	0.392	-
3.9567	3110	0.4011	-
3.9695	3120	0.3946	-
3.9822	3130	0.3883	-
3.9949	3140	0.3939	-
4.0	3144	-	0.2901
4.0076	3150	0.3879	-
4.0204	3160	0.3909	-
4.0331	3170	0.3975	-
4.0458	3180	0.3987	-
4.0585	3190	0.3919	-
4.0712	3200	0.4138	-
4.0840	3210	0.382	-
4.0967	3220	0.4063	-
4.1094	3230	0.3885	-
4.1221	3240	0.4015	-
4.1349	3250	0.4007	-
4.1476	3260	0.4045	-
4.1603	3270	0.4094	-
4.1730	3280	0.386	-
4.1858	3290	0.4054	-
4.1985	3300	0.3927	-
4.2112	3310	0.3996	-
4.2239	3320	0.4042	-
4.2366	3330	0.3954	-
4.2494	3340	0.3827	-
4.2621	3350	0.4022	-
4.2748	3360	0.4044	-
4.2875	3370	0.3988	-
4.3003	3380	0.4062	-
4.3130	3390	0.4005	-
4.3257	3400	0.4067	-
4.3384	3410	0.3825	-
4.3511	3420	0.3878	-
4.3639	3430	0.4012	-
4.3766	3440	0.3966	-
4.3893	3450	0.3871	-
4.4020	3460	0.3998	-
4.4148	3470	0.4033	-
4.4275	3480	0.4069	-
4.4402	3490	0.3994	-
4.4529	3500	0.3969	-
4.4656	3510	0.398	-
4.4784	3520	0.3945	-
4.4911	3530	0.3884	-
4.5	3537	-	0.2890
4.5038	3540	0.4028	-
4.5165	3550	0.3852	-
4.5293	3560	0.3986	-
4.5420	3570	0.3942	-
4.5547	3580	0.4043	-
4.5674	3590	0.4064	-
4.5802	3600	0.3904	-
4.5929	3610	0.3934	-
4.6056	3620	0.3861	-
4.6183	3630	0.3927	-
4.6310	3640	0.4085	-
4.6438	3650	0.3951	-
4.6565	3660	0.408	-
4.6692	3670	0.3902	-
4.6819	3680	0.3966	-
4.6947	3690	0.4102	-
4.7074	3700	0.3961	-
4.7201	3710	0.4022	-
4.7328	3720	0.4056	-
4.7455	3730	0.4078	-
4.7583	3740	0.3974	-
4.7710	3750	0.3961	-
4.7837	3760	0.3861	-
4.7964	3770	0.3933	-
4.8092	3780	0.3902	-
4.8219	3790	0.3938	-
4.8346	3800	0.3913	-
4.8473	3810	0.3985	-
4.8601	3820	0.4055	-
4.8728	3830	0.3933	-
4.8855	3840	0.3968	-
4.8982	3850	0.3959	-
4.9109	3860	0.3933	-
4.9237	3870	0.4057	-
4.9364	3880	0.4023	-
4.9491	3890	0.3938	-
4.9618	3900	0.3969	-
4.9746	3910	0.4019	-
4.9873	3920	0.3922	-
5.0	3930	0.3955	0.2884
5.0127	3940	0.3968	-
5.0254	3950	0.3899	-
5.0382	3960	0.3861	-
5.0509	3970	0.3866	-
5.0636	3980	0.4033	-
5.0763	3990	0.4031	-
5.0891	4000	0.4154	-
5.1018	4010	0.4002	-
5.1145	4020	0.4006	-
5.1272	4030	0.3959	-
5.1399	4040	0.4043	-
5.1527	4050	0.3904	-
5.1654	4060	0.3939	-
5.1781	4070	0.3824	-
5.1908	4080	0.3945	-
5.2036	4090	0.3897	-
5.2163	4100	0.3983	-
5.2290	4110	0.3925	-
5.2417	4120	0.4056	-
5.2545	4130	0.4121	-
5.2672	4140	0.4013	-
5.2799	4150	0.4016	-
5.2926	4160	0.3944	-
5.3053	4170	0.393	-
5.3181	4180	0.3999	-
5.3308	4190	0.3833	-
5.3435	4200	0.3719	-
5.3562	4210	0.3934	-
5.3690	4220	0.4009	-
5.3817	4230	0.3843	-
5.3944	4240	0.4003	-
5.4071	4250	0.3971	-
5.4198	4260	0.399	-
5.4326	4270	0.4037	-
5.4453	4280	0.3965	-
5.4580	4290	0.3849	-
5.4707	4300	0.4083	-
5.4835	4310	0.4102	-
5.4962	4320	0.3944	-
5.5	4323	-	0.2881
5.5089	4330	0.4029	-
5.5216	4340	0.3967	-
5.5344	4350	0.3874	-
5.5471	4360	0.3922	-
5.5598	4370	0.3903	-
5.5725	4380	0.402	-
5.5852	4390	0.4023	-
5.5980	4400	0.4039	-
5.6107	4410	0.3954	-
5.6234	4420	0.4044	-
5.6361	4430	0.4067	-
5.6489	4440	0.391	-
5.6616	4450	0.3968	-
5.6743	4460	0.4016	-
5.6870	4470	0.4007	-
5.6997	4480	0.3893	-
5.7125	4490	0.3978	-
5.7252	4500	0.3981	-
5.7379	4510	0.3908	-
5.7506	4520	0.3952	-
5.7634	4530	0.4013	-
5.7761	4540	0.3892	-
5.7888	4550	0.4012	-
5.8015	4560	0.3881	-
5.8142	4570	0.3799	-
5.8270	4580	0.3933	-
5.8397	4590	0.3787	-
5.8524	4600	0.3886	-
5.8651	4610	0.4028	-
5.8779	4620	0.4022	-
5.8906	4630	0.4038	-
5.9033	4640	0.3886	-
5.9160	4650	0.3981	-
5.9288	4660	0.3884	-
5.9415	4670	0.3992	-
5.9542	4680	0.4065	-
5.9669	4690	0.3918	-
5.9796	4700	0.4078	-
5.9924	4710	0.3928	-
6.0	4716	-	0.2879
6.0051	4720	0.3939	-
6.0178	4730	0.4158	-
6.0305	4740	0.4007	-
6.0433	4750	0.3929	-
6.0560	4760	0.3977	-
6.0687	4770	0.3993	-
6.0814	4780	0.3776	-
6.0941	4790	0.391	-
6.1069	4800	0.3894	-
6.1196	4810	0.397	-
6.1323	4820	0.4013	-
6.1450	4830	0.3871	-
6.1578	4840	0.3901	-
6.1705	4850	0.3958	-
6.1832	4860	0.3978	-
6.1959	4870	0.385	-
6.2087	4880	0.3959	-
6.2214	4890	0.3897	-
6.2341	4900	0.399	-
6.2468	4910	0.3987	-
6.2595	4920	0.3971	-
6.2723	4930	0.3914	-
6.2850	4940	0.4009	-
6.2977	4950	0.3919	-
6.3104	4960	0.3933	-
6.3232	4970	0.3919	-
6.3359	4980	0.3761	-
6.3486	4990	0.3865	-
6.3613	5000	0.3816	-
6.3740	5010	0.3898	-
6.3868	5020	0.3919	-
6.3995	5030	0.3908	-
6.4122	5040	0.4045	-
6.4249	5050	0.4044	-
6.4377	5060	0.3916	-
6.4504	5070	0.3954	-
6.4631	5080	0.4123	-
6.4758	5090	0.4027	-
6.4885	5100	0.403	-
6.5	5109	-	0.2877
6.5013	5110	0.4009	-
6.5140	5120	0.3901	-
6.5267	5130	0.4065	-
6.5394	5140	0.4033	-
6.5522	5150	0.3908	-
6.5649	5160	0.3996	-
6.5776	5170	0.4002	-
6.5903	5180	0.3884	-
6.6031	5190	0.399	-
6.6158	5200	0.3921	-
6.6285	5210	0.3913	-
6.6412	5220	0.3926	-
6.6539	5230	0.3929	-
6.6667	5240	0.3965	-
6.6794	5250	0.3944	-
6.6921	5260	0.4034	-
6.7048	5270	0.3867	-
6.7176	5280	0.4031	-
6.7303	5290	0.3926	-
6.7430	5300	0.3923	-
6.7557	5310	0.3977	-
6.7684	5320	0.3987	-
6.7812	5330	0.4027	-
6.7939	5340	0.3995	-
6.8066	5350	0.405	-
6.8193	5360	0.4113	-
6.8321	5370	0.3974	-
6.8448	5380	0.4014	-
6.8575	5390	0.3972	-
6.8702	5400	0.3953	-
6.8830	5410	0.3992	-
6.8957	5420	0.393	-
6.9084	5430	0.3865	-
6.9211	5440	0.3892	-
6.9338	5450	0.3955	-
6.9466	5460	0.3954	-
6.9593	5470	0.4041	-
6.9720	5480	0.3927	-
6.9847	5490	0.416	-
6.9975	5500	0.3895	-
7.0	5502	-	0.2877
7.0102	5510	0.3927	-
7.0229	5520	0.3896	-
7.0356	5530	0.3926	-
7.0483	5540	0.3907	-
7.0611	5550	0.3916	-
7.0738	5560	0.399	-
7.0865	5570	0.3882	-
7.0992	5580	0.4003	-
7.1120	5590	0.3928	-
7.1247	5600	0.4001	-
7.1374	5610	0.4098	-
7.1501	5620	0.3948	-
7.1628	5630	0.3988	-
7.1756	5640	0.3893	-
7.1883	5650	0.3905	-
7.2010	5660	0.4024	-
7.2137	5670	0.4011	-
7.2265	5680	0.3959	-
7.2392	5690	0.4014	-
7.2519	5700	0.3932	-
7.2646	5710	0.385	-
7.2774	5720	0.3964	-
7.2901	5730	0.3922	-
7.3028	5740	0.3864	-
7.3155	5750	0.3864	-
7.3282	5760	0.4018	-
7.3410	5770	0.406	-
7.3537	5780	0.3975	-
7.3664	5790	0.405	-
7.3791	5800	0.3872	-
7.3919	5810	0.3885	-
7.4046	5820	0.3942	-
7.4173	5830	0.3913	-
7.4300	5840	0.41	-
7.4427	5850	0.3957	-
7.4555	5860	0.4052	-
7.4682	5870	0.4082	-
7.4809	5880	0.3889	-
7.4936	5890	0.4008	-
7.5	5895	-	0.2876
7.5064	5900	0.4027	-
7.5191	5910	0.4029	-
7.5318	5920	0.398	-
7.5445	5930	0.3914	-
7.5573	5940	0.3943	-
7.5700	5950	0.3983	-
7.5827	5960	0.3955	-
7.5954	5970	0.3984	-
7.6081	5980	0.3987	-
7.6209	5990	0.391	-
7.6336	6000	0.4049	-
7.6463	6010	0.3837	-
7.6590	6020	0.3899	-
7.6718	6030	0.3945	-
7.6845	6040	0.398	-
7.6972	6050	0.3902	-
7.7099	6060	0.3926	-
7.7226	6070	0.4101	-
7.7354	6080	0.3832	-
7.7481	6090	0.3985	-
7.7608	6100	0.3847	-
7.7735	6110	0.3905	-
7.7863	6120	0.3756	-
7.7990	6130	0.3921	-
7.8117	6140	0.3906	-
7.8244	6150	0.397	-
7.8372	6160	0.3971	-
7.8499	6170	0.4037	-
7.8626	6180	0.3873	-
7.8753	6190	0.3871	-
7.8880	6200	0.3917	-
7.9008	6210	0.3843	-
7.9135	6220	0.4039	-
7.9262	6230	0.4004	-
7.9389	6240	0.4038	-
7.9517	6250	0.3908	-
7.9644	6260	0.4022	-
7.9771	6270	0.4061	-
7.9898	6280	0.3799	-
8.0	6288	-	0.2875

Framework Versions

• Python: 3.11.13
• Sentence Transformers: 5.1.2
• Transformers: 4.57.1
• PyTorch: 2.8.0+cu129
• Accelerate: 1.11.0
• Datasets: 4.3.0
• 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",
}

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}
}