This is a sentence-transformers model finetuned from microsoft/mpnet-base 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: microsoft/mpnet-base
Maximum Sequence Length: 512 tokens
Output Dimensionality: 768 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- json

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': 'MPNetModel'})
  (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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("sahithkumar7/final-mpnet-base-fullfinetuned-epoch5")
# Run inference
sentences = [
    'What is the number of genes obtained from comparing control and LIPUS-stimulated samples?',
    'Differentially expressed genes (DEGs) were obtained\nbetween control and LIPUS-stimulated samples using\nan adjusted P<0.05 and|log2FC| > 1 as cutoffs to define\nstatistically significant differential expression. 676 genes\nwere obtained from which 578 were upregulated when\nstimulated with LIPUS and 98 genes were subregulated\n(Supp. Figure 1). To further understand the functions\nand pathways associated with the differentially expressed\ngenes (DEG), Gene Ontology (GO) and Kyoto Encyclo-\npedia of Genes and Genomes (KEGG) analyses were con-\nducted using the DAVID database [37, 38].',
    'independent studies have shown a raising trend in both cancer incidence [2] and a high-salt\ndietary lifestyle [7], there is no direct correlation between dietary salt intake and breast\ncancer. Interestingly, in the human body, certain organs such as the skin and lymph nodes\nhave a natural tendency to accumulate salt [8]. Although unknown, the pathophysiological\nsignificance of this selective accumulation of sodium in certain organs and solid tumors is\nan area of intense research.',
]
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,  0.6568, -0.0120],
#         [ 0.6568,  1.0000, -0.0343],
#         [-0.0120, -0.0343,  1.0000]])

Evaluation

Metrics

Triplet

Datasets: initial_test and final_test
Evaluated with TripletEvaluator

Metric	initial_test	final_test
cosine_accuracy	0.98	0.98

Training Details

Training Dataset

json

Dataset: json
Size: 800 training samples
Columns: anchor, positive, and negative
Approximate statistics based on the first 800 samples:
anchor positive negative
type string string string
details
min: 7 tokens
mean: 16.79 tokens
max: 39 tokens
min: 39 tokens
mean: 117.74 tokens
max: 265 tokens
min: 40 tokens
mean: 116.14 tokens
max: 356 tokens

Samples:

anchor	positive	negative
`What is the limitation of FBG-based sensors in tactile feedback?`	Furthermore, FBG-based 3-axis tactile sensors have been[object Object]proposed for a more comprehensive haptic perception tool[object Object]in surgeries (Figure 1D) (16). Five optical fibers merged[object Object]with FBG sensors are suspended in a deformable medium[object Object]and measure the compression or tension of the tissue as the[object Object]sensors are pressed against it, returning a _ surface[object Object]reaction map. While FBG-based sensors are small, flexible, and[object Object]sensitive, there are several challenges that need to be[object Object]addressed for optimal performance for tactile feedback. These[object Object]sensors are temperature sensitive, requiring temperature	141]. Therefore, it is not known to what extent spared[object Object]axons are remyelinated by transplanted Schwann cells,[object Object]nor is the contribution of this myelin to functional im-[object Object]provements proven. Transplantation of Schwann cells[object Object]incapable of producing myelin, such as cells derived[object Object]from trembler (Pmp22Tr) mutant mice, may be useful[object Object]in establishing a causal relationship between myelin re-[object Object]generation and functional improvements. Several MSC[object Object]transplantations demonstrate an increase of myelin re-[object Object]tention and the number of myelinated axons in the le-[object Object]sion site during a chronic post-injury period [57]. Thus,
`What are the advantages of strain elastography?`	frontiersin.org[object Object][object Object]--- Page 8 ---[object Object]Kumar et al.[object Object][object Object]TABLE 2 Modalities of ultrasound elastography.[object Object][object Object]Modality[object Object]Strain elastography[object Object][object Object]Excitation[object Object]Applied manual compression (38)[object Object][object Object]Advantages[object Object][object Object]No additional specialized equipment[object Object]required (40)[object Object][object Object]10.3389/fmedt.2023.1238129[object Object][object Object]Limitations[object Object][object Object]Qualitative measurements (39)[object Object][object Object]Internal physiological mechanism (42)[object Object][object Object]Simple low-cost design (40)[object Object][object Object]Applied compression is operator-dependent (51)[object Object][object Object]More commonly used (52)[object Object][object Object]High inter-observer variability (51)[object Object][object Object]coustic radiation force impulse Acoustic radiation force (43)[object Object][object Object](ARFI) imaging[object Object][object Object]Image beyond slip boundaries (45)	Publisher’s Note: MDPI stays neutral[object Object]with regard to jurisdictional claims in[object Object]published maps and institutional afil-[object Object][object Object]iations.[object Object][object Object]onon)[object Object][object Object]Copyright: © 2021 by the author.[object Object]Licensee MDPI, Basel, Switzerland.[object Object]This article is an open access article[object Object]distributed under the terms and[object Object]conditions of the Creative Commons[object Object]Attribution (CC BY) license (https://[object Object]creativecommons.org/licenses/by/[object Object]4.0/).[object Object][object Object]Joan and Sanford I. Weill Department of Medicine, Weill Cornell Medical College, 525 East 68th Street,[object Object]Room M-522, Box 130, New York, NY 10065, USA; [object Object] or [object Object]
`What is the material used for the substrate in a piezoelectric element?`	gain for biomedical applications.[object Object][object Object]frontiersin.org[object Object][object Object]--- Page 9 ---[object Object]Kumar et al.[object Object][object Object]>[object Object][object Object][PMUT ][object Object][object Object]Electrode: Voltage Electrode2[object Object][object Object]© piezoelectric elements[object Object]o[object Object][object Object]—: OSi02[object Object][object Object]©) silicon substrate[object Object][object Object]B [ CMUT ][object Object]AC DC[object Object][object Object]membrane[object Object][object Object]—————[object Object][object Object]vacuum[object Object]insulator[object Object][object Object]substrate[object Object][object Object]= ground[object Object][object Object]FIGURE 3	Histopatholo[object Object]Cytology Total, n (%) Benign, n (%) P ey Cancer, n (%)[object Object]FA 2 (15.4%) FTC 2 (25%)[object Object]0 GD (7.7%) PTC 6 (75%)[object Object]I 21 (4.0%) NG 9 (69.2%)[object Object]Other diagnosis (7.7%)[object Object]FA 15 (9.9%) FIC 4 (14.3%)[object Object]FT-UMP (0.7%) MTC 3 (10.7%)[object Object]GD (0.7%) PTC 21 (75%)[object Object]Il 180 (34.5%) OA (0.7%)[object Object]LT (0.7%)[object Object]NG 130 (85.5%)[object Object]NIFTP 2 (1.3%)[object Object]FA 14 (23.7%) FIC 7 (28.0%)[object Object]FI-UMP 2 (3.4%) OTC 1 (4.0%)[object Object]OA (1.7%) PTC 17 (68.0%)[object Object]Il 84 (16.1%) LT 3 (5.1%)[object Object]NG 35 (59.3%)[object Object]NIFTP 2 (3.4%)[object Object]WDT-UMP 2 (3.4%)[object Object]FA 15 (26.3%) OTC 1 (7.7%)[object Object]FT-UMP 5 (8.8%) PTC 12 (92.3%)[object Object]OA 13 (22.8%)[object Object]IV 70 (13.4%) LT 2 (3.5%)[object Object]NG 18 (31.6%)[object Object]NIFTP 2 (3.5%)

Loss: MultipleNegativesRankingLoss with these parameters:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim" 4}

Evaluation Dataset

json

Dataset: json
Size: 200 evaluation samples
Columns: anchor, positive, and negative
Approximate statistics based on the first 200 samples:
anchor positive negative
type string string string
details
min: 7 tokens
mean: 17.14 tokens
max: 35 tokens
min: 40 tokens
mean: 121.3 tokens
max: 356 tokens
min: 45 tokens
mean: 119.75 tokens
max: 356 tokens

Samples:

anchor	positive	negative
`What can differentiate into a very wide variety of tissues?`	lead to decreased rates of graft-versus-host disease. They[object Object]also can differentiate into a very wide variety of tissues. For[object Object]example, when compared with bone marrow stem cells or[object Object]mobilized peripheral blood, umbilical cord blood stem cells[object Object]have a greater repopulating ability.5° Cord blood derived[object Object]CD34+ cells have very potent hematopoietic abilities, and[object Object]this is attributed to the immaturity of the stem cells rela-[object Object]tive to adult derived cells. Studies have been done that an-[object Object]alyze long term survival of children with hematologic dis-[object Object]orders who were transplanted with umbilical cord blood	metabolic regulation may affect the function of more than one organelle. Therefore, if the[object Object]miR-17-92 regulatory cluster can perturb genes related to mitochondrial metabolic function,[object Object]it could be also related, in some way, to genes involved in lysosomal metabolic function.[object Object]Lysosomes are intracellular organelles that, in form of small vesicles, participate in[object Object]several cellular functions, mainly digestion, but also vesicle trafficking, autophagy, nutrient[object Object]sensing, cellular growth, signaling [85], and even enzyme secretion. The membrane-bound
`What are the two most common types of pluripotent stem cells?`	III]. AMNIOTIC CELLS AS A SOURCE FOR STEM[object Object]CELLS[object Object][object Object]Historically, the two most common types of pluripotent[object Object]stem cells include embryonic stem cells (ESCs) and induced[object Object]pluripotent stem cells (iPSCs).35 However, despite the many[object Object]research efforts to improve ESC and iPSC technologies,[object Object]there are still enormous clinical challenges.°> Two signif-[object Object]icant issues posed by ESC and iPSC technologies include[object Object]low survival rate of transplanted cells and tumorigenicity.°>[object Object]Recently, researchers have isolated pluripotent stem cells	Explanation: criterion 6 indicates a positive diagnosis only within the DC VI group[object Object]relative to all other categories. Criterion 5 indicates a positive diagnosis within the DCs VI[object Object]and V relative to all other categories.[object Object][object Object]The highest positive predictive value (PPV) confirming malignancy through histopatho-[object Object]logical examination for criterion 6 was 0.93, and for criterion 5, it was 0.92. For the subsequent[object Object]criteria, the PPVs were as follows: criterion 4—0.66; criterion 3—0.55; criterion 2—0.40.
`What percentage of stem cells are present in bone marrow?`	`ing 30% in some tissues.43-45 This is a significant difference[object Object]from the .0001-.0002% stem cells present in bone marrow.43[object Object]Given this difference in stem cell concentration between[object Object]the sources, there will be more ADSCs per sample of WAT`	migration of bCSCs. This finding raises the possibil-[object Object]ity that LIPUS may decrease the ability of these cells to[object Object]invade adjacent tissues and start the process of metasta-[object Object]ses. These results also suggested that some of the changes[object Object]induced by LIPUS take longer to be detected in this type[object Object]of 2D migration model, possible due to changes in gene[object Object]expression pattern. To further study this hypothesis, we[object Object]performed a Transwell invasion assay. The data revealed[object Object]a reduced number of cells crossing the membrane after[object Object]LIPUS stimulation, indicating that therapeutic LIPUS

Loss: MultipleNegativesRankingLoss with these parameters:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim" 4}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
num_train_epochs: 5
warmup_ratio: 0.1
fp16: True
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 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: 1.0
num_train_epochs: 5
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
use_ipex: 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}
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: 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
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
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Click to expand

Epoch	Step	Training Loss	Validation Loss	initial_test_cosine_accuracy	final_test_cosine_accuracy
-1	-1	-	-	0.7800	-
0.02	1	3.124	-	-	-
0.04	2	3.2227	-	-	-
0.06	3	3.1108	-	-	-
0.08	4	3.1317	-	-	-
0.1	5	3.334	-	-	-
0.12	6	3.0113	-	-	-
0.14	7	3.1324	-	-	-
0.16	8	2.9737	-	-	-
0.18	9	3.06	-	-	-
0.2	10	3.0135	-	-	-
0.22	11	2.6313	-	-	-
0.24	12	2.6668	-	-	-
0.26	13	3.0152	-	-	-
0.28	14	2.2725	-	-	-
0.3	15	2.0777	-	-	-
0.32	16	2.1797	-	-	-
0.34	17	1.5419	-	-	-
0.36	18	1.5243	-	-	-
0.38	19	1.9736	-	-	-
0.4	20	1.3789	1.1273	0.9200	-
0.42	21	1.3214	-	-	-
0.44	22	1.1415	-	-	-
0.46	23	1.2763	-	-	-
0.48	24	1.2017	-	-	-
0.5	25	0.8564	-	-	-
0.52	26	1.1946	-	-	-
0.54	27	0.6691	-	-	-
0.56	28	1.0806	-	-	-
0.58	29	0.9866	-	-	-
0.6	30	0.6018	-	-	-
0.62	31	0.7234	-	-	-
0.64	32	1.1117	-	-	-
0.66	33	0.9061	-	-	-
0.68	34	0.6709	-	-	-
0.7	35	0.3719	-	-	-
0.72	36	0.4776	-	-	-
0.74	37	0.7804	-	-	-
0.76	38	0.6527	-	-	-
0.78	39	0.4496	-	-	-
0.8	40	0.4754	0.6063	0.9800	-
0.82	41	0.5374	-	-	-
0.84	42	0.4347	-	-	-
0.86	43	0.2719	-	-	-
0.88	44	0.4964	-	-	-
0.9	45	0.5999	-	-	-
0.92	46	0.3679	-	-	-
0.94	47	1.1593	-	-	-
0.96	48	0.4503	-	-	-
0.98	49	0.7891	-	-	-
1.0	50	0.7262	-	-	-
1.02	51	0.384	-	-	-
1.04	52	0.2659	-	-	-
1.06	53	0.2753	-	-	-
1.08	54	0.7633	-	-	-
1.1	55	0.3969	-	-	-
1.12	56	0.7593	-	-	-
1.1400	57	0.4218	-	-	-
1.16	58	0.5748	-	-	-
1.18	59	0.247	-	-	-
1.2	60	0.2258	0.6012	0.9800	-
1.22	61	0.3377	-	-	-
1.24	62	0.2809	-	-	-
1.26	63	0.1531	-	-	-
1.28	64	0.3147	-	-	-
1.3	65	0.1778	-	-	-
1.32	66	0.5214	-	-	-
1.34	67	0.3154	-	-	-
1.3600	68	0.3562	-	-	-
1.38	69	0.0939	-	-	-
1.4	70	0.1446	-	-	-
1.42	71	0.1135	-	-	-
1.44	72	0.1814	-	-	-
1.46	73	0.2597	-	-	-
1.48	74	0.3724	-	-	-
1.5	75	0.1345	-	-	-
1.52	76	0.7102	-	-	-
1.54	77	0.2938	-	-	-
1.56	78	0.5879	-	-	-
1.58	79	0.6798	-	-	-
1.6	80	0.2806	0.4559	1.0	-
1.62	81	0.1723	-	-	-
1.6400	82	0.3122	-	-	-
1.6600	83	0.2249	-	-	-
1.6800	84	0.2339	-	-	-
1.7	85	0.664	-	-	-
1.72	86	0.4421	-	-	-
1.74	87	0.3913	-	-	-
1.76	88	0.1642	-	-	-
1.78	89	0.2313	-	-	-
1.8	90	0.1561	-	-	-
1.8200	91	0.3084	-	-	-
1.8400	92	0.6486	-	-	-
1.8600	93	0.1822	-	-	-
1.88	94	0.3486	-	-	-
1.9	95	0.414	-	-	-
1.92	96	0.3011	-	-	-
1.94	97	0.2638	-	-	-
1.96	98	0.2688	-	-	-
1.98	99	0.1398	-	-	-
2.0	100	0.2447	0.4312	0.9600	-
2.02	101	0.2479	-	-	-
2.04	102	0.3743	-	-	-
2.06	103	0.1115	-	-	-
2.08	104	0.158	-	-	-
2.1	105	0.161	-	-	-
2.12	106	0.1981	-	-	-
2.14	107	0.0744	-	-	-
2.16	108	0.0901	-	-	-
2.18	109	0.0332	-	-	-
2.2	110	0.0482	-	-	-
2.22	111	0.1093	-	-	-
2.24	112	0.0683	-	-	-
2.26	113	0.0923	-	-	-
2.2800	114	0.0389	-	-	-
2.3	115	0.0404	-	-	-
2.32	116	0.0354	-	-	-
2.34	117	0.0424	-	-	-
2.36	118	0.1524	-	-	-
2.38	119	0.0868	-	-	-
2.4	120	0.047	0.4226	0.9800	-
2.42	121	0.046	-	-	-
2.44	122	0.0249	-	-	-
2.46	123	0.0332	-	-	-
2.48	124	0.0582	-	-	-
2.5	125	0.0288	-	-	-
2.52	126	0.0835	-	-	-
2.54	127	0.0132	-	-	-
2.56	128	0.0562	-	-	-
2.58	129	0.0382	-	-	-
2.6	130	0.0419	-	-	-
2.62	131	0.0296	-	-	-
2.64	132	0.0909	-	-	-
2.66	133	0.2664	-	-	-
2.68	134	0.0778	-	-	-
2.7	135	0.0646	-	-	-
2.7200	136	0.0429	-	-	-
2.74	137	0.1977	-	-	-
2.76	138	0.1503	-	-	-
2.7800	139	0.0582	-	-	-
2.8	140	0.1864	0.4420	0.9800	-
2.82	141	0.0845	-	-	-
2.84	142	0.0359	-	-	-
2.86	143	0.0684	-	-	-
2.88	144	0.0375	-	-	-
2.9	145	0.2441	-	-	-
2.92	146	0.0728	-	-	-
2.94	147	0.0423	-	-	-
2.96	148	0.0745	-	-	-
2.98	149	0.1564	-	-	-
3.0	150	0.0416	-	-	-
3.02	151	0.0121	-	-	-
3.04	152	0.052	-	-	-
3.06	153	0.0255	-	-	-
3.08	154	0.1191	-	-	-
3.1	155	0.0104	-	-	-
3.12	156	0.0257	-	-	-
3.14	157	0.027	-	-	-
3.16	158	0.0305	-	-	-
3.18	159	0.0292	-	-	-
3.2	160	0.0194	0.4015	0.9600	-
3.22	161	0.018	-	-	-
3.24	162	0.0153	-	-	-
3.26	163	0.0318	-	-	-
3.2800	164	0.0221	-	-	-
3.3	165	0.0087	-	-	-
3.32	166	0.0458	-	-	-
3.34	167	0.1314	-	-	-
3.36	168	0.0511	-	-	-
3.38	169	0.1193	-	-	-
3.4	170	0.1112	-	-	-
3.42	171	0.0646	-	-	-
3.44	172	0.1594	-	-	-
3.46	173	0.022	-	-	-
3.48	174	0.0097	-	-	-
3.5	175	0.0904	-	-	-
3.52	176	0.0142	-	-	-
3.54	177	0.0433	-	-	-
3.56	178	0.032	-	-	-
3.58	179	0.025	-	-	-
3.6	180	0.02	0.4191	0.9800	-
3.62	181	0.0629	-	-	-
3.64	182	0.0269	-	-	-
3.66	183	0.0207	-	-	-
3.68	184	0.0319	-	-	-
3.7	185	0.0401	-	-	-
3.7200	186	0.0078	-	-	-
3.74	187	0.0347	-	-	-
3.76	188	0.1676	-	-	-
3.7800	189	0.0089	-	-	-
3.8	190	0.0096	-	-	-
3.82	191	0.0257	-	-	-
3.84	192	0.0176	-	-	-
3.86	193	0.0084	-	-	-
3.88	194	0.0068	-	-	-
3.9	195	0.0188	-	-	-
3.92	196	0.0535	-	-	-
3.94	197	0.0142	-	-	-
3.96	198	0.0153	-	-	-
3.98	199	0.096	-	-	-
4.0	200	0.0084	0.3991	0.9800	-
4.02	201	0.0081	-	-	-
4.04	202	0.0061	-	-	-
4.06	203	0.0081	-	-	-
4.08	204	0.0507	-	-	-
4.1	205	0.013	-	-	-
4.12	206	0.0081	-	-	-
4.14	207	0.0577	-	-	-
4.16	208	0.0114	-	-	-
4.18	209	0.0049	-	-	-
4.2	210	0.0044	-	-	-
4.22	211	0.0662	-	-	-
4.24	212	0.0288	-	-	-
4.26	213	0.0162	-	-	-
4.28	214	0.0653	-	-	-
4.3	215	0.0078	-	-	-
4.32	216	0.021	-	-	-
4.34	217	0.0915	-	-	-
4.36	218	0.0135	-	-	-
4.38	219	0.0754	-	-	-
4.4	220	0.007	0.4001	0.9800	-
4.42	221	0.0061	-	-	-
4.44	222	0.0169	-	-	-
4.46	223	0.004	-	-	-
4.48	224	0.0175	-	-	-
4.5	225	0.0068	-	-	-
4.52	226	0.0073	-	-	-
4.54	227	0.0127	-	-	-
4.5600	228	0.0106	-	-	-
4.58	229	0.0118	-	-	-
4.6	230	0.0455	-	-	-
4.62	231	0.0094	-	-	-
4.64	232	0.0583	-	-	-
4.66	233	0.0267	-	-	-
4.68	234	0.0085	-	-	-
4.7	235	0.0203	-	-	-
4.72	236	0.0089	-	-	-
4.74	237	0.1229	-	-	-
4.76	238	0.0092	-	-	-
4.78	239	0.0218	-	-	-
4.8	240	0.0526	0.4114	0.9800	-
4.82	241	0.0163	-	-	-
4.84	242	0.0295	-	-	-
4.86	243	0.0197	-	-	-
4.88	244	0.0022	-	-	-
4.9	245	0.0174	-	-	-
4.92	246	0.0223	-	-	-
4.9400	247	0.0074	-	-	-
4.96	248	0.0115	-	-	-
4.98	249	0.0228	-	-	-
5.0	250	0.0381	-	-	-
-1	-1	-	-	-	0.9800

Framework Versions

Python: 3.11.13
Sentence Transformers: 5.0.0
Transformers: 4.52.4
PyTorch: 2.6.0+cu124
Accelerate: 1.8.1
Datasets: 3.6.0
Tokenizers: 0.21.2

Citation

BibTeX

Sentence Transformers

bibtex

@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

bibtex

@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}
}

	anchor	positive	negative
type	string	string	string
details	min: 7 tokens mean: 16.79 tokens max: 39 tokens	min: 39 tokens mean: 117.74 tokens max: 265 tokens	min: 40 tokens mean: 116.14 tokens max: 356 tokens