trentmkelly/qwen3-embedding-0.6b-stylometry-v9

SentenceTransformer

This is a sentence-transformers model trained on the json dataset. It maps sentences & paragraphs to a 1024-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: Unknown -->
• —Maximum Sequence Length: 512 tokens
• —Output Dimensionality: 1024 dimensions
• —Similarity Function: Cosine Similarity
• —Training Dataset:
• —json <!-- - Language: Unknown --> <!-- - License: Unknown -->

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}) with Transformer model: Qwen3Model 
  (1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': True, '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 = [
    'El documento del PSOE donde habla de "personas de sexo biológico masculino" tiene un lenguaje cuya finalidad es decir "las mujeres trans en realidad no son mujeres". Recuerdo cuando sacaron a la luz datos médicos personales de Imane Khelif, violando sus derechos a la privacidad y calificandola así. || La ley trans de Argentina, la primera en el mundo en garantizar la autodeterminación del género. Ahora mismo en peligro. || @subetealanutria.bsky.social  Arrestaron a tu nutria. www.elcomercio.es/aviles/polic... || Lo anacrónicas que son ya las monarquías, y el olor a rancio que desprenden. || Yo no quiero que está guerra me deshumanice. No soy nacionalista, ni tampoco sionista. Simplemente soy judía, y ni siquiera la religión es un aspecto tan importante en mi vida. Tengo amigas árabes a las que quiero conservar como tales. Mi pareja es católica. Creo en la convivencia y en el respeto. || Cuenten conmigo para ello. Muchos judíos lo decimos, no en nuestro nombre. Como bien dices, la religión de cada quien es algo personal. Y el estado, las leyes, etc... tienen que ser neutrales en ese sentido. || En el pensamiento judío, históricamente hay un concepto, el "Kol Israel arevim ze laze" que expresa una suerte de responsabilidad colectiva del pueblo judío. Entendida en que cada judío es responsable el uno por el otro. Es algo que muchos no compartimos, porque cada persona es responsable por sí. || Soy trans y vaya que he perdido amigos, hasta lazos familiares se han roto. Pero es que, la transfobia es tan grande en la sociedad, que yo he notado incluso como mi género llega a afectar a mi pareja. La gente la juzga todo el tiempo, y es que como somos dos mujeres, hay quien no nos entiende. || Esos son como la canción de Luis Miguel: "la incondicional, la que no espera nada". || En EEUU en lugar de un doctor de prestigio al timón de la Secretaría de Salud, vamos a tener a un antivacunas. En Texas han incorporado estudios bíblicos creacionistas al currículo escolar. A cargo de la defensa, nada de un militar de carrera, a un presentador de la Fox que no sabe donde esta España',
    'Exactamente, o sea, como una enseñanza religiosa de los tiempos de Moisés, pues está bien. Pero no podemos hacer leyes basadas en responsabilidades colectivas, o en preceptos religiosos. El problema es que en Israel, los haredim cada vez tienen más peso e influencia política. Es un acoso total. || A mi me decían cosas horribles sobre los árabes. Racismo puro, pero mi espíritu es rebelde, y por eso quise conocer más sobre ellos. Cuando empecé a conocer chicas árabes, entendí que esas ideas no tenían sentido. Descubrí una cultura muy hospitalaria y que guarda mucha relación con la nuestra. || ???????? EE.UU: Viernes Negro (Black Friday) para seguir con la fiebre consumista. || Desde aquel último día de Sucot, cuando empezó esta última guerra, he vivido una crisis personal. Soy judía y mi idea sobre Israel ha pasado de cierta simpatía a cuestionarme su legitimidad. Estoy harta de tanta violencia. De hecho le he expresado mis sentimientos de afecto a todos mis amigos árabes || Le pide a un pueblo mestizo que celebre un genocidio de indígenas. Con lo que nos ha costado dejar de celebrar el dichoso Día de la Raza y darle otro significado. En fin, como dijo Madero, de tal palo tal astilla. || No sean mal pensados. Quiere hacer un pastel de calabaza y sorprender a su novia/o. Lo demás es por si surge plan. || Hay gente que gracias a sus padrinos, están por encima de la ley. Ya hay países como Francia, Alemania y Hungría que han dicho que no van a detener a Netanyahu. Como signatarios del Estatuto de Roma, tienen la obligación de detenerlo, porque sobre él pesa una orden de arresto por crímenes de guerra. || Vi sheyne ir kuk. Her Majesty. Que guapa se mira usted, Su Majestad. || Pues eso sería con la ley romana, porque lo que es la judía literalmente se la saltó a la torera. || Yo siento que soy más bien una señora estilo Toulouse Lautrec. || Apalizan a enfermeras, cuya vocación es cuidar a las personas enfermas. Cuando simplemente están pidiendo un salario justo y condiciones laborales dignas. Es una vergüenza. Cuanta gente votó a Milei y se ha dado cuenta ya a la mala como les cagaron.',
    'Yaaaaa ????. Lo importante es el alma, el sitio y sus pobladores. No le podemos quitar mérito a quienes dan todo aquí desde el principio sólo porque no les gustan los mismos nombres para denominar a un post, que es lo que son realmente. Yo soy emotiva, pero poco dada a palabras cursis ????, me cuestan || Gracias, QQ! || Muuuuuuuu malo. || Espero que te ayude el médico ???????????????? || Es que eso es lo que me parece lo normal. No se me había ocurrido que alguien fuera a mirar las cifras de seguidores de alguien antes de dar me gusta o RT a un post. Me quedo loquer. || Jajajaja. He cenado hace una hora. Creo que tiempo insuficiente para asentar lo ingerido ???? || Gracias por hacerme verlo así ???? || Ayyy ???? ???? || Ay el gif ???? || Ahhh jajaja. Pues es cuando hablas contando un tema que se refiere a una persona concreta, pero sin mencionarla. Como el chiste de Gila del señor pasando junto a un asesino y diciendo al aire “alguien ha matado a alguien… alguien es un asesino…”. || Gracias!! || Desvelada por tercer día consecutivo. Tengo unos picores insoportables en la piel que no me dejan dormir y me despierto. || Oii, qué bonito ???? || ???????? || Una conversación de machos discutiendo sobre si las mujeres hacen bien o mal el feminismo, pasando ahora por mis narices, no, que me enamoro. || Además, alguna vez recibí una medio queja de quien no se veía en el circulito, como si fuera culpa mía que salga alguien o no, como si los hiciéramos a mano ????????\u200d♀️ || Justo lo que yo pensaba ???? || Aaaaaanda narices! || Ay, Gracias Mery ???? || El problema es que si sigues a 5.000, ya no puedes leerlas. Precisamente porqué me gusta leer e interactuar sin perderme mucho, no concibo poder seguir a 400 personas. Acabaría teniendo un batiburrillo en el que encuentro a los amigos. Seguir a 5000 cuentas creo que es como un tener un tío en Murcia || Es que lo idea sería no saber ni preocuparse por saber cuántos seguidores tiene una cuenta. ¿Te gusta un tuit? Pues RT. No me entra en la cabeza otro uso. || Felicidades!! || 3. El paritorio. || Pero si justo aquí no hay ninguna pseudo, solo una mención como un piano ???? || Muchas Gracias Yoto ????',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

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

<!--

Direct Usage (Transformers)

<details><summary>Click to see the direct usage in Transformers</summary>

</details> -->

<!--

Downstream Usage (Sentence Transformers)

You can finetune this model on your own dataset.

<details><summary>Click to expand</summary>

</details> -->

<!--

Out-of-Scope Use

List how the model may foreseeably be misused and address what users ought not to do with the model. -->

Evaluation

Metrics

Binary Classification

• —Dataset: test_pairs
• —Evaluated with <code>BinaryClassificationEvaluator</code>

<!--

Bias, Risks and Limitations

What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. -->

<!--

Recommendations

What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. -->

Training Details

Training Dataset

json

• —Dataset: json
• —Size: 438,990 training samples
• —Columns: <code>anchor</code>, <code>positive</code>, and <code>negative</code>
• —Approximate statistics based on the first 1000 samples: | | anchor | positive | negative | |:--------|:--------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------|:--------------------------------------------------------------------------------------| | type | string | string | string | | details | <ul><li>min: 398 tokens</li><li>mean: 503.44 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 392 tokens</li><li>mean: 503.36 tokens</li><li>max: 512 tokens</li></ul> | <ul><li>min: 407 tokens</li><li>mean: 505.06 tokens</li><li>max: 512 tokens</li></ul> |
• —Samples: | anchor | positive | negative | |:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------| | <code>The validation they're looking for is showing they can make us mad. Give them contempt instead. Never sound hurt or offended; they interpret that as, they have exerted power over us. So, give them something they DON'T want. Contempt works well, or insult their genital size. || On 10/7, Gaza chose war. They're getting war, all the war they could want and then some. I don't see the problem. If you assholes ever actually want peace - IF - you'll demand that Gaza stop firing rockets at Israel, return the hostages, and accept a ceasefire. Until then you are full of shit. || The reason we need pointers is, when you pass parameters to a function - in C and some other languages anyway - there is no way to pass those parameters back when they change. Pointers are the clumsy workaround: you pass a LOCATION to the function, and then work with the data at that location. || They don't like that they're losing, but I don't think many of them have realized they're just getting what they deserve for...</code> | <code>Popper's paradox there, it doesn't even strike me as a paradox so much as a bad-faith understanding of tolerance. We believe in tolerance only until it starts actually threatening others; then we have to restrict it. But up to that point, tolerance harms no one and is the right idea. || It's almost the season: www.youtube.com/watch?v=HYFQ... || You're blaming Garland for what Cannon and the Supreme Court have done. On this fine Thanksgiving you may fuck all the way off and just keep going. || That map is riddled with errors. That's not Saudi Arabia, it's Louisiana. || Damn. This is how Daleks should always look, I think I would welcome a Dalek invasion. || That is nearly pornographically good. || Trick question, there is no starvation in Gaza, but you antisemitic dumbasses will believe any propaganda against Israel. Try following an account like this one to see what life in Gaza is actually like. Real videos posted by real Gazans. bsky.app/profile/imsh... || Israel is the sort of co...</code> | <code>Need not be an “army” to engage in war as hamasses has. Somewhere around 37800 gazans and 1700 Israelis estimated based on the last reported numbers. The disparity in numbers is largely not relevant || Ok money || And you take this guy seriously ?! He has never been to the areas he is talking about to see if indeed this is true. This is just bullshit built on top of bullshit. || They do but very few. Many many socialists. || hizbollasses predictably broke the ceasefire today. || Correction: The curse of letting homocidal maniacs parading around as religious pious people run amok in your country. || It doesn’t deter them because they have been brainwashed by the islamists to value death over life. || “Many”.. if you found many as being in the low 1000s and mostly part of a fringe of society group. || Oh please. You think you have the market cornered on memes ?! Are you 2y old?! Facts are what matters. Phony accusations of genocide only cheapen your stance. || Then the WB should build in...</code> | | <code>My gift to you babe???????? || Wouldn’t think that a person who clearly on drugs should be the head of the FBI || I knew what you meant! ???? || How things have changed! || There could have 4 years of poverty & joy! || Sometimes it’s better not knowing the answer, it’s better to repair the damage & make sure it doesn’t happen again! || Well done Lori Jo || If BlueSky keeps growing how it has over the last 2 weeks, in 6 months X will be in trouble! || It didn’t work in the 80’s & it won’t work now! If anything it probably encouraged people to use drugs at the time. || Compulsory voting works! || No family dinner || I believe I’ve seen a pattern in these bot accounts, I believe they are computer generated so the way they set up to are all the same! || 20 years it ran for! 1955 to 75! || @desilydic.bsky.social Welcome to BlueSky Desi || Have a good day Shannon, don’t physically harm any family Trunts! I need you here! || Not another one || Moms know their Sons || All over the world ???? ...</code> | <code>Followed a couple new to me || He should be worried, his investors have been known to use bone saws! || Get rid of them! || Sure to follow back True Blue, make smart comments & reposts, she’s a much larger account than mind & can grow your audience quicker! || Yep, one thing about information technology, your here one day and you’re gone the next! || It’s the female scammers that are the problem, the magrats are easy to spot. || I did use follow you on X, before Elon threw me off 9 months ago! || Followed almost everyone || That’s a great analogy! || Well done Diane || Thanks for the follow back Amy Hope you had a great night as well, || Yep! || I only did economics to year 9 at secondary school, I learn why Tariffs & Subsidies were bad for the economy, it was kind of reinforced by the teachers! || Hi Lori Jo || I’ve got a name for you, @paulatilbrook.bsky.social She’s from WA, she could do with some reposting etc, if you don’t follow her already. || @jenny-algies56.bsky.social Thanks...</code> | <code>???????????? || I try to confuse the bots for funsies. || Thank you for correctly identifying it as DISinformation. || Some dude in NJ is dumber than fuck. || It's quite alright and something I very much understand. || How else would you stab him? || Awesome! You're welcome. || Following based on your word. || Highly recommend the MAGA cockroaches block list. These were already all captured. || Ooh ooh me please! ????????‍♂️ || Botox || Wait you were a wrestler?! || If you lost 200+ lbs you've accomplished more than I ever have so feel good about yourself! || Clearly it's all about feeding fake women to involuntary incels who can't be around real women. || I'm laying in bed binging Yellowstone! Zero fucks || Porn bots Lottery winners Africans begging for money Illuminati-devotees Finance scammers And on and on || For the second tweet, the irony is dripping. Look no further than Felon Schmuck for how to be a grade school douchebag. || What an awesome picture || Time has remained a dece...</code> | | <code>Questo è sicuro: ma la rubrica di Dan Savage è imperdibile! || Voi GenZ siete dei dandies, altroché. || Dici bene grossa. Quindi implicitamente se non sono viventi le definisci non-specie. A meno che per te un'entità che per mettere in croce 100 parole a pappagallo consuma l'equivalente energetico di 3 litri d'acqua sia condizione sufficiente a definirla vivente. || Sono diventate una specie? Sono viventi? || Maledetti drughè!!! Festa di Halloween abusiva con 600 partecipanti, denunciati gli organizzatori www.trevisotoday.it/cronaca/sant... > L'evento nella serata dello scorso 31 ottobre a Santa Lucia di Piave negli spazi del ristorante "Al Gabbiano". Nei guai un uomo e una donna, rispettivamente... || Non ha ancora 'menzionato' Paolo Borsellino (tipo Pasolini con la pula)? || La fasciocrazia di Meloni e le altre notizie di MicroMega www.micromega.net/la-fasciocra... > Ripercorriamo insieme la settimana di MicroMega (25-29 novembre) attraverso i contributi più salienti pubblicati ...</code> | <code>È morta Peggy Caserta, amica (e amante?) di Janis Joplin, di cui ha raccontato la vita privata | Rolling Stone Italia www.rollingstone.it/musica/news-... > Aveva 84 anni. Ha scritto della rockstar in due libri, ‘Going Down With Janis’ (in parte ripudiato) e ‘I Ran Into Some Trouble’. Ha cont... || Tenerezza || Come ho fatto fin'adesso??? @posillipostore.bsky.social for President! || Ma non solo... Pensa alle mafie, pensa a chi non ha padronanza della lingua - sempreché abbia disponibilità dei documenti, etc etc. Anche se il mestiere è lo stesso un conto è una libera scelta, un conto la schiavitù. ... || Ah meno male. Anche a me tutti sti binari mi fanno paura. Se vogliono i binari, che vadano alla stazione no? || ... Sono 2 universi in 2 dimensioni parallele. E poi aggiungiamo maschi vs. femmine: è un frattale. Mi brucia pensarlo, figurati dirlo e scriverlo pubblicamente: ma anche no, grazie. || Chi c’è dietro Bluesky? Tutto su Jay Graber, l’attivista digitale a capo del social n...</code> | <code>Oggi sciopero generale e black friday. #ossimori || Comunque parti (divertiti!) ???? || Nulla si crea, nulla si distrugge (e W il djin' 4 evah!) www.youtube.com/watch?v=Fpao... || Quella sempre! Beata te, dove vai? || Gratis! || ???????????? || Già, come eravamo (siamo) ridicoli || (Così ha più senso) || (Sentita la prima puntata poco fa..) || Se facessi lo sceneggiatore di serie sci-fi e tenessi un taccuino sul comodino su cui segnare i miei sogni, avrei materiale per anni || Stanotte ho sognato che mio cognato era super ricco, alto 10 cm in più, vestito trendy e veniva a casa mia con mia sorella (io in bermuda e giubbotto pesante); il mio forno diventava lavatrice piena di panni e non riuscivamo ad aprirlo. Poi tutti in stazione a Bologna ad aspettare il treno! ???? || Io l’ho impostata su “solo complimenti” || Un amico mi ha appena chiesto di lavorare con lui a Capodanno (ho passato talmente tanti capodanni in disco a mettere musica e a far divertire gli altri che mi stava scappan...</code> |
• —Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• —per_device_train_batch_size: 128
• —per_device_eval_batch_size: 128
• —gradient_accumulation_steps: 2
• —learning_rate: 2e-05
• —weight_decay: 0.01
• —num_train_epochs: 1
• —warmup_ratio: 0.05
• —bf16: True
• —gradient_checkpointing: True
• —batch_sampler: no_duplicates

All Hyperparameters

<details><summary>Click to expand</summary>

• —overwrite_output_dir: False
• —do_predict: False
• —eval_strategy: no
• —prediction_loss_only: True
• —per_device_train_batch_size: 128
• —per_device_eval_batch_size: 128
• —per_gpu_train_batch_size: None
• —per_gpu_eval_batch_size: None
• —gradient_accumulation_steps: 2
• —eval_accumulation_steps: None
• —torch_empty_cache_steps: None
• —learning_rate: 2e-05
• —weight_decay: 0.01
• —adam_beta1: 0.9
• —adam_beta2: 0.999
• —adam_epsilon: 1e-08
• —max_grad_norm: 1.0
• —num_train_epochs: 1
• —max_steps: -1
• —lr_scheduler_type: linear
• —lr_scheduler_kwargs: {}
• —warmup_ratio: 0.05
• —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: 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: {'minnumparams': 0, 'xla': False, 'xlafsdpv2': False, 'xlafsdpgrad_ckpt': False}
• —fsdp_transformer_layer_cls_to_wrap: None
• —accelerator_config: {'splitbatches': False, 'dispatchbatches': None, 'evenbatches': True, 'useseedablesampler': True, 'nonblocking': False, 'gradientaccumulationkwargs': None}
• —parallelism_config: None
• —deepspeed: None
• —label_smoothing_factor: 0.0
• —optim: adamw_torch
• —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: True
• —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: no_duplicates
• —multi_dataset_batch_sampler: proportional

</details>

Training Logs

Framework Versions

• —Python: 3.11.12
• —Sentence Transformers: 3.3.1
• —Transformers: 4.57.0
• —PyTorch: 2.7.0+cu128
• —Accelerate: 1.12.0
• —Datasets: 4.8.4
• —Tokenizers: 0.22.2

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}
}

<!--

Glossary

Clearly define terms in order to be accessible across audiences. -->

<!--

Model Card Authors

Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. -->

<!--

Model Card Contact

Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->