metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:32394
  - loss:MultipleNegativesRankingLoss
base_model: sentence-transformers/all-MiniLM-L6-v2
widget:
  - source_sentence: >-
      Feel-good Mexican telenovelas set in the 1980s with a focus on elementary
      school kids and their relationships.
    sentences:
      - >-
        Title: America: The Story of Us

        Genres: Documentary

        Overview: From wagon trains crossing the untamed frontier to man's first
        steps on the moon, this series offers a compelling look at the people,
        inventions and events that helped forge the United States of America.

        Tagline: 

        Creator: 

        Stars: Liev Schreiber, Tom Brokaw, Annette Gordon-Reed

        Release Date: 2010-04-25

        Keywords: history
      - >-
        Title: Carrusel

        Genres: Comedy, Drama, Family, Kids, Soap

        Overview: Carrusel is a Mexican telenovela, produced by and first
        broadcast on Televisa in 1989. It covers daily life in a Mexican
        elementary school and the children's relationships with a charismatic
        teacher named Jimena. Among other plot devices, it deals with the
        differences between the upper and lower classes of Mexican society —
        specifically as seen in a romantic relationship between Cirilo, a poor
        black boy, and a spoiled rich girl, Maria Joaquina Villaseñor.

        Tagline: 

        Creator: Abel Santacruz

        Stars: Gabriela Rivero, Pedro Javier Viveros, Ludwika Paleta

        Release Date: 1989-01-19

        Keywords: mexico city, mexico, elementary school, school, family, naive
        children, 1980s, school kids
      - >-
        Title: Dracula

        Genres: Drama

        Overview: It's the late 19th century, and the mysterious Dracula has
        arrived in London, posing as an American entrepreneur who wants to bring
        modern science to Victorian society. He's especially interested in the
        new technology of electricity, which promises to brighten the night -
        useful for someone who avoids the sun. But he has another reason for his
        travels: he hopes to take revenge on those who cursed him with
        immortality centuries earlier. Everything seems to be going according to
        plan... until he becomes infatuated with a woman who appears to be a
        reincarnation of his dead wife.

        Tagline: The legend takes new life.

        Creator: Daniel Knauf, Cole Haddon

        Stars: Jonathan Rhys Meyers, Jessica De Gouw, Katie McGrath

        Release Date: 2013-10-25

        Keywords: london, england, vampire, victorian england, 19th century,
        dracula
  - source_sentence: Know any good TV programs with both Lee Dong-wook and Yoo In-na?
    sentences:
      - >-
        Title: Touch Your Heart

        Genres: Comedy, Drama

        Overview: Hoping to make a comeback after a bad scandal, an actress
        agrees to research a new role by taking a job as a secretary for a
        prickly attorney.

        Tagline: 

        Creator: Park Joon-hwa

        Stars: Lee Dong-wook, Yoo In-na, Lee Sang-woo

        Release Date: 2019-02-06

        Keywords: based on novel or book, assistant, romance, lawyer, law firm,
        opposites attract, entertainment industry, famous actress
      - >-
        Title: Creeped Out

        Genres: Sci-Fi & Fantasy, Mystery

        Overview: A masked figure known as "The Curious" collects tales of dark
        magic, otherworldly encounters and twisted technology in this kids
        anthology series.

        Tagline: 

        Creator: Robert Butler, Bede Blake

        Stars: Aurora Aksnes, William Romain, Jaiden Cannatelli

        Release Date: 2017-10-31

        Keywords: anthology, horror anthology, horror
      - >-
        Title: Love a Lifetime

        Genres: Drama, Sci-Fi & Fantasy, Action & Adventure

        Overview: Amidst a legacy of family feuds, a kind-hearted young woman,
        Rong Hua, crosses paths with the mysterious Nalan Yue while searching
        for a powerful healing artifact. As they fall in love, they uncover a
        deep history of revenge linking their families. With a new threat rising
        and Nalan Yue battling a dark power within, the two must fight to
        overcome the past and protect their future together.

        Tagline: 

        Creator: 

        Stars: Ren Jialun, Zhang Huiwen, Li Yitong

        Release Date: 2020-06-18

        Keywords: love at first sight, romance, hatred, wuxia, successor, web
        series, secondary couple
  - source_sentence: >-
      Memorable drama TV programs focused on life and grappling with
      relationships
    sentences:
      - |-
        Title: El Maleficio
        Genres: Drama
        Overview: 
        Tagline: 
        Creator: Fernanda Villeli
        Stars: Fernando Colunga, Marlene Favela, Sofía Castro
        Release Date: 2023-11-13
        Keywords: 
      - >-
        Title: You Can Do Better

        Genres: Comedy

        Overview: A half-hour brain candy show that tackles major topics like
        drinking, technology, sex, money, and friends. Through a mix of sketch,
        how-to, man-on-the-street and expert interviews, our hosts impart tips
        and tricks that every adult should know. Viewers will learn to be better
        at the subjects no one teaches in school, and they'll get to belly-laugh
        along the way.

        Tagline: 

        Creator: 

        Stars: Abbi Crutchfield, Matthew Latkiewicz, Jessy Greer

        Release Date: 2016-08-23

        Keywords: 
      - >-
        Title: Junjou Romantica

        Genres: Animation, Comedy, Drama

        Overview: Three couples, three intense romances: a student’s tutor
        crosses the line, a loner meets a force of nature, and a carefree man
        faces love he can’t ignore.

        Tagline: 

        Creator: Shungiku Nakamura

        Stars: Hikaru Hanada, Takahiro Sakurai, Nobutoshi Canna

        Release Date: 2008-04-10

        Keywords: college, romance, slice of life, coming of age, based on
        manga, art, teacher student relationship, lgbt, angst, anime, drastic
        change of life, erotic, gay theme, tsundere, boys' love (bl)
  - source_sentence: Compelling dramas exploring the repercussions of past actions
    sentences:
      - >-
        Title: Stay Close

        Genres: Drama, Crime, Mystery

        Overview: When Carlton Flynn vanishes 17 years to the night after
        Stewart Green did, it sets off a chain reaction in the lives of people
        connected to both men.

        Tagline: Everyone has secrets.

        Creator: Harlan Coben

        Stars: Cush Jumbo, James Nesbitt, Richard Armitage

        Release Date: 2021-12-31

        Keywords: suicide, detective, celebrity, reporter, husband, dark
      - |-
        Title: Los misterios de Laura
        Genres: Crime, Drama, Mystery
        Overview: 
        Tagline: 
        Creator: Javier Holgado, Carlos Vila
        Stars: María Pujalte, Fernando Guillén Cuervo, César Camino
        Release Date: 2009-07-27
        Keywords: investigation, investigator, crime investigation
      - >-
        Title: Hitori no Shita: The Outcast

        Genres: Animation, Sci-Fi & Fantasy, Action & Adventure, Comedy

        Overview: Zhang Chulan leads a very common college student's life until
        he finds himself caught up in a terrible incident that happened in a
        small village. As he was walking through a graveyard, he is assaulted by
        zombies. Thinking that it was over for him, a mysterious girl carrying a
        sword suddenly saves him and disappears.

        Tagline: 

        Creator: Dong Man Tang, Mi Er

        Stars: Xiao Liansha, Sheng Feng, Yuntu Cao

        Release Date: 2016-07-09

        Keywords: fighting, advanture, city, based on manhua, fantasy, urban
        fantasy, sino japanese production, passionate, donghua, comedy,
        coproduction, urban adventure, qihuan, dongfang
  - source_sentence: >-
      Memorable drama TV series focused on slight romance and grappling with
      investigation
    sentences:
      - >-
        Title: Reset

        Genres: Drama, Mystery

        Overview: The lives of a college student and a video game designer are
        kept being reset after an explosion on a bus. During each reset, they
        have to work together to find out what the reason for the explosion is.
        Will these two be able to save themselves and their fellow passengers?
        Will they be able to close the time-loop?

        Tagline: 

        Creator: 

        Stars: Bai Jingting, Zhao Jinmai, Liu Tao

        Release Date: 2022-01-11

        Keywords: time travel, investigation, time loop, explosion, slight
        romance, student, suspense
      - >-
        Title: The Boss

        Genres: Comedy, Drama

        Overview: Eliseo is the superintendent of an upscale building. On the
        surface, is cordial and docile in his role, but underneath Eliseo
        believes himself the omnipotent figure of the community — meddling in
        the affairs of residents and pulling strings as he sees fit. Eliseo's
        only concern is protecting his job, which comes under threat by a
        proposed pool project.

        Tagline: 

        Creator: Mariano Cohn, Gastón Duprat

        Stars: Gastón Cocchiarale, Guillermo Francella, Gabriel Goity

        Release Date: 2022-10-26

        Keywords: manipulation, buenos aires, argentina, apartment building,
        scheming, serie argentina, building superintendent
      - >-
        Title: Wildlife Specials

        Genres: Documentary

        Overview: The BBC Wildlife Specials are a series of nature documentary
        programmes commissioned by BBC Television. The Wildlife Specials began
        with a pilot episode in 1995. 20 programmes have been made to date, with
        three of the recent ones being in multi parts. The earlier programmes
        were produced in-house by the BBC's specialist Natural History Unit, but
        the more recent Spy in the... titles were made by the independent John
        Downer Productions. The first 18 programmes, up to 2008, were narrated
        by David Attenborough. The most recent two were narrated by David
        Tennant.


        "The world's leading natural history filmmakers meet the world's most
        charismatic animals"


        — BBC tagline

        Tagline: 

        Creator: 

        Stars: David Attenborough

        Release Date: 1995-04-14

        Keywords: animals, nature documentary, cats
pipeline_tag: sentence-similarity
library_name: sentence-transformers

SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2. It maps sentences & paragraphs to a 384-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: sentence-transformers/all-MiniLM-L6-v2
Maximum Sequence Length: 256 tokens
Output Dimensionality: 384 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': 256, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 384, '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): Normalize()
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("sentence_transformers_model_id")
# Run inference
sentences = [
    'Memorable drama TV series focused on slight romance and grappling with investigation',
    'Title: Reset\nGenres: Drama, Mystery\nOverview: The lives of a college student and a video game designer are kept being reset after an explosion on a bus. During each reset, they have to work together to find out what the reason for the explosion is. Will these two be able to save themselves and their fellow passengers? Will they be able to close the time-loop?\nTagline: \nCreator: \nStars: Bai Jingting, Zhao Jinmai, Liu Tao\nRelease Date: 2022-01-11\nKeywords: time travel, investigation, time loop, explosion, slight romance, student, suspense',
    "Title: The Boss\nGenres: Comedy, Drama\nOverview: Eliseo is the superintendent of an upscale building. On the surface, is cordial and docile in his role, but underneath Eliseo believes himself the omnipotent figure of the community — meddling in the affairs of residents and pulling strings as he sees fit. Eliseo's only concern is protecting his job, which comes under threat by a proposed pool project.\nTagline: \nCreator: Mariano Cohn, Gastón Duprat\nStars: Gastón Cocchiarale, Guillermo Francella, Gabriel Goity\nRelease Date: 2022-10-26\nKeywords: manipulation, buenos aires, argentina, apartment building, scheming, serie argentina, building superintendent",
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

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

Training Details

Training Dataset

Unnamed Dataset

Size: 32,394 training samples
Columns: sentence_0, sentence_1, and sentence_2

Approximate statistics based on the first 1000 samples:

	sentence_0	sentence_1	sentence_2
type	string	string	string
details	min: 8 tokens mean: 17.07 tokens max: 38 tokens	min: 38 tokens mean: 133.54 tokens max: 256 tokens	min: 40 tokens mean: 132.86 tokens max: 256 tokens

Samples:

sentence_0	sentence_1	sentence_2
`Dramatic fantasy romance with a touch of destiny and betrayal`	Title: Eternal Love Genres: Drama, Sci-Fi & Fantasy Overview: Three hundred years ago, Bai Qian stood on the Zhu Xian Terrace, turned around and jumped off without regret. Ye Hua stood by the bronze mirror to witness with his own eyes her death. Three hundred years later, in the East Sea Dragon Palace, the two meet unexpectedly. Another lifetime another world, after suffering betrayal Bai Qian no longer feels anything, yet she can't seem to comprehend Ye Hua's actions. Three lives three worlds, her and him, are they fated to love again? Tagline: Creator: Stars: Yang Mi, Mark Chao, Ken Chang Tzu-Yao Release Date: 2017-01-30 Keywords: china, arranged marriage, romance, fate, second chance, older woman younger man relationship, xianxia	Title: Kidsongs Genres: Comedy Overview: Kidsongs is an American children's media franchise which includes Kidsongs Music Video Stories on DVD and video, The Kidsongs TV Show, CDs of favorite children’s songs and covers of oldies and pop hits from the 50s, 60s and 70s, song books, sheet music, toys and an ecommerce website. Kidsongs was created by producer/writer Carol Rosenstein and director Bruce Gowers of Together Again Video Productions, both of whom are music video and television production veterans. The duo had produced and directed over 100 music videos for Warner Brothers Records and took their idea of music videos for children to the record label. Warner Brothers funded the first video, “A Day at Old MacDonald’s Farm”. Shortly thereafter, a three way partnership between TAVP, WBR and View-Master Video was formed with TAVP being responsible for production and WBR and View-Master responsible for distribution to video and music stores, and toy stores respectively. Tagline: Creat...
`Memorable animation TV shows focused on cartoon and grappling with superliga`	`Title: Supa Strikas Genres: Animation Overview: With dreams of becoming Super League champions, a talented striker named Shakes and his football team take on rivals while going on global adventures. Tagline: Creator: Stars: Corny Rempel, Kevin Aichele, Chelsea Rankin Release Date: 2009-02-15 Keywords: cartoon, football (soccer), superliga`	`Title: Grand Hotel Genres: Drama, Crime, Mystery Overview: Santiago Mendoza owns last family-owned hotel in multicultural Miami Beach, while his glamorous second wife, Gigi, and their adult children enjoy the spoils of success. Tagline: Five star hotel. Five star secrets. Creator: Brian Tanen Stars: Demián Bichir, Roselyn Sánchez, Denyse Tontz Release Date: 2019-06-17 Keywords: miami, florida, hotel, remake, family conflict, upstairs downstairs, wealthy family`
`Any recommendations for top action & adventure TV programs from 2010 featuring Catherine Siachoque?`	Title: Missing Genres: Mystery, Action & Adventure, Crime Overview: The night Elisa’s cousins-Santiago, Flor, and Eduardo, invited her to a nightclub and after a great deal of begging her parents allowed her go. When Danna and he sister-in-law Cecilia went to pick them up, all of them started showing up except for Elisa. As the hours passed, her parent grew more and more desperate and it was then when they decided to call the police and file a missing report. Tagline: Creator: Stars: Sonya Smith, Catherine Siachoque, Jesus Licciardello Release Date: 2010-03-08 Keywords:	`Title: Aurora Genres: Mystery, Soap, Drama, Crime Overview: Having been cryogenically frozen for 20 years, Aurora's heart torn between past and present : memories of an old love and chance of a new one. Tagline: Creator: Marcela Citterio Stars: Sara Maldonado, Eugenio Siller, Sonya Smith Release Date: 2010-11-01 Keywords:`

Loss: MultipleNegativesRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 32
per_device_eval_batch_size: 32
num_train_epochs: 4
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: 32
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: 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: 4
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}
tp_size: 0
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: batch_sampler
multi_dataset_batch_sampler: round_robin

Training Logs

Epoch	Step	Training Loss
0.4936	500	0.864
0.9872	1000	0.5835
1.4808	1500	0.4604
1.9743	2000	0.4476
2.4679	2500	0.3866
2.9615	3000	0.3688
3.4551	3500	0.3353
3.9487	4000	0.3385

Framework Versions

Python: 3.11.12
Sentence Transformers: 3.4.1
Transformers: 4.51.3
PyTorch: 2.6.0+cu124
Accelerate: 1.6.0
Datasets: 3.5.1
Tokenizers: 0.21.1

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}
}