/ Модель

Характеристики

Контекст
8 194 токена
Параметры
659 млн
Архитектура
xlm-roberta · dense
Базовая модель
bge-reranker-v2-m3
Библиотека
sentence-transformers
Опубликована
27 июн 2026

bge-reranker-v2-m3-amazon

Ранжирование текстаТекст → score9 профессий

Многоязычный кросс-энкодер для оценки релевантности пар текстов и переранжирования результатов поиска.

19скачиваний за 30 дней0лайковвеса обновлены27 июн 2026

Бенчмарки

Как модель справилась с задачами пользователей ailist и что о ней публикуют открытые лидерборды

Тесты пользователейпока нет

Никто ещё не опубликовал тест этой модели. Соберите бенчмарк из своих задач — результат появится здесь.

  1. Добавьте модель в бенчмарк

    Вместе с другими моделями категории «Ранжирование» — сравнение идёт внутри одной категории.

  2. Опишите свои задачи

    Напишите запросы, при необходимости приложите файлы и эталонные ответы.

  3. Запустите и опубликуйте

    Модели ответят на каждую задачу, судья оценит ответы. Опубликованный результат появится на этой странице.

Публичные оценки

Из открытых лидербордов. Сравнивайте результаты одного теста при одинаковых условиях — отсутствие оценки не значит, что модель хуже.

В подключённых источниках результатов для этой модели пока нет.

Что умеет модель

Описание составлено по карточке модели и метаданным Hugging Face

Это кросс-энкодерный переранжировщик steerrec, дообученный на базе BAAI/bge-reranker-v2-m3 с библиотекой Sentence Transformers. Он оценивает пары текстов одним числовым баллом, который можно использовать для переранжирования результатов поиска и семантического поиска; в README приведена оценка на наборах amazon_test и amazon_val. Модель имеет архитектуру XLM-RoBERTa, содержит 658,7 млн параметров и принимает последовательности длиной до 512 токенов по README; метаданные каталога указывают контекст до 8194 токенов. Лицензия в предоставленных данных не указана.

Сценарии применения

  • переранжирование результатов поиска по запросу
  • оценка релевантности документов запросу
  • улучшение порядка результатов семантического поиска
  • повторная сортировка товаров по текстовому запросу

Ограничения

README заявляет максимальную длину последовательности 512 токенов, тогда как метаданные каталога указывают контекст 8194 токена; ориентируйтесь на ограничение используемой конфигурации. Модель оценивает пары текстов и сама по себе не генерирует ответы.

Для каких профессий подходит9

Разделы каталога профессий, в задачах которых используется эта модель

Все профессии

Также подходит для

Карточка модели

README от автора на Hugging Face, как он его опубликовал

Открыть на Hugging Face

CrossEncoder based on BAAI/bge-reranker-v2-m3

This is a Cross Encoder model finetuned from BAAI/bge-reranker-v2-m3 using the sentence-transformers library. It computes scores for pairs of texts, which can be used for text reranking and semantic search.

Model Details

Model Description

  • Model Type: Cross Encoder
  • Base model: BAAI/bge-reranker-v2-m3
  • Maximum Sequence Length: 512 tokens
  • Number of Output Labels: 1 label

Model Sources

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 CrossEncoder

# Download from the 🤗 Hub
model = CrossEncoder("steerrec/bge-reranker-v2-m3-amazon")
# Get scores for pairs of texts
pairs = [
    ["I want something emotional and heartfelt, but not boring. I don't want another slow, dull show.", 'Civil Love Product description Civil Love Review Every so often you come across a movie you really enjoy. For me, this was it. This film contains everything a good movie needs: a good story, fine acting, direction, and a purpose. The theme of this film is that we are too quick to judge people sometimes and that people can occasionally surprise us. --Edwin L. Carpenter, Dove Foundation Movies & TV Genre for Featured Categories Drama'],
    ["I want something emotional and heartfelt, but not boring. I don't want another slow, dull show.", 'Surprised by Love A young businesswoman tries to convince her uptight parents to accept her boyfriend, but then she falls for an old high-school flame. Stars Hilarie Burton, Paul Campbell, and Tim Conway. IMDb 6.7 1 h 24 min 2015 7+ Drama Cheerful Passionate'],
    ["I want something emotional and heartfelt, but not boring. I don't want another slow, dull show.", 'All I Want for Christmas Tis the season for making dreams true. If everything goes according to plan for one young boy, he just may be sharing this Christmas with a new dad. Ever since her husband died, Sarah Armstrong has divided her time between working at a community center and caring for her nine-year-old son, Jesse. When Jesse wins the All I Want for Christmas contest sponsored by a toy company, his wish comes as no surprise to find a new dad. Just in time for the holidays, All I Want for Christmas is a surprise romantic treat for anyone who believes in the magic of love, friendship, and family. Movies & TV Holidays & Seasonal Christmas'],
    ["I just finished a few really touching movies and shows that left me feeling deeply moved and emotionally invested. I'm now looking for something that keeps that same heartfelt, character-driven vibe but with a bit more depth and complexity. I want to watch something that's a mix of drama and drama, maybe with some drama or drama, but it's hard to say. I'm looking for something that combines the elements of a story, a character, and the setting of the characters. It's like a movie where the plot is driven by the people involved, not just the events. I'm craving something that's not too fast-paced, but still has enough substance to keep me engaged. Maybe something that's a little more complex than your average film, but not so much that it feels overwhelming. I'm also interested in something that's not too dark, but still has some emotional weight. I want to feel connected to the characters and their journeys, even if they're not the most straightforward. I'm in the mood for something that's a bit more than just a movie. I need something that's a bit more intricate, with a strong narrative that's not just about the characters but the world they inhabit. I'm looking for a show that's not just a collection of characters but a full, immersive experience. I'm not sure what exactly I", "The Guardian Action-packed drama that takes you inside the never-before-seen world of the elite Coast Guard Rescue Swimmers. Reeling with grief from a mission gone tragically wrong, a legendary swimmer is given a mission he doesn't want -- training raw recruits. IMDb 6.9 2 h 19 min 2006 X-Ray PG-13 Drama Adventure Compelling Inspiring"],
    ["I just finished a few really touching movies and shows that left me feeling deeply moved and emotionally invested. I'm now looking for something that keeps that same heartfelt, character-driven vibe but with a bit more depth and complexity. I want to watch something that's a mix of drama and drama, maybe with some drama or drama, but it's hard to say. I'm looking for something that combines the elements of a story, a character, and the setting of the characters. It's like a movie where the plot is driven by the people involved, not just the events. I'm craving something that's not too fast-paced, but still has enough substance to keep me engaged. Maybe something that's a little more complex than your average film, but not so much that it feels overwhelming. I'm also interested in something that's not too dark, but still has some emotional weight. I want to feel connected to the characters and their journeys, even if they're not the most straightforward. I'm in the mood for something that's a bit more than just a movie. I need something that's a bit more intricate, with a strong narrative that's not just about the characters but the world they inhabit. I'm looking for a show that's not just a collection of characters but a full, immersive experience. I'm not sure what exactly I", 'Magnolia [VHS] An intriguing and entertaining study in characters going through varying levels of crisis and introspection. This psychological drama leads you in several different directions, weaving and intersecting various subplots and characters, from a brilliant Tom Cruise, as a self-proclaimed pied-piper, to a child forced to go on a TV game show and the pressures he faces from a ruthless father. Movies & TV Genre for Featured Categories Drama'],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

# Or rank different texts based on similarity to a single text
ranks = model.rank(
    "I want something emotional and heartfelt, but not boring. I don't want another slow, dull show.",
    [
        'Civil Love Product description Civil Love Review Every so often you come across a movie you really enjoy. For me, this was it. This film contains everything a good movie needs: a good story, fine acting, direction, and a purpose. The theme of this film is that we are too quick to judge people sometimes and that people can occasionally surprise us. --Edwin L. Carpenter, Dove Foundation Movies & TV Genre for Featured Categories Drama',
        'Surprised by Love A young businesswoman tries to convince her uptight parents to accept her boyfriend, but then she falls for an old high-school flame. Stars Hilarie Burton, Paul Campbell, and Tim Conway. IMDb 6.7 1 h 24 min 2015 7+ Drama Cheerful Passionate',
        'All I Want for Christmas Tis the season for making dreams true. If everything goes according to plan for one young boy, he just may be sharing this Christmas with a new dad. Ever since her husband died, Sarah Armstrong has divided her time between working at a community center and caring for her nine-year-old son, Jesse. When Jesse wins the All I Want for Christmas contest sponsored by a toy company, his wish comes as no surprise to find a new dad. Just in time for the holidays, All I Want for Christmas is a surprise romantic treat for anyone who believes in the magic of love, friendship, and family. Movies & TV Holidays & Seasonal Christmas',
        "The Guardian Action-packed drama that takes you inside the never-before-seen world of the elite Coast Guard Rescue Swimmers. Reeling with grief from a mission gone tragically wrong, a legendary swimmer is given a mission he doesn't want -- training raw recruits. IMDb 6.9 2 h 19 min 2006 X-Ray PG-13 Drama Adventure Compelling Inspiring",
        'Magnolia [VHS] An intriguing and entertaining study in characters going through varying levels of crisis and introspection. This psychological drama leads you in several different directions, weaving and intersecting various subplots and characters, from a brilliant Tom Cruise, as a self-proclaimed pied-piper, to a child forced to go on a TV game show and the pressures he faces from a ruthless father. Movies & TV Genre for Featured Categories Drama',
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

Evaluation

Metrics

Cross Encoder Reranking

Metricamazon_testamazon_val
map0.85180.8329
mrr@100.88880.8814
ndcg@100.90860.9007

Training Details

Training Dataset

Unnamed Dataset

  • Size: 320,937 training samples
  • Columns: sentence1, sentence2, and label
  • Approximate statistics based on the first 1000 samples:
    sentence1sentence2label
    typestringstringfloat
    detailsmin: 78 charactersmean: 537.26 charactersmax: 1341 charactersmin: 48 charactersmean: 1079.61 charactersmax: 10645 charactersmin: 0.0mean: 0.33max: 1.0
  • Samples:
    sentence1sentence2label
    I want something emotional and heartfelt, but not boring. I don't want another slow, dull show.Civil Love Product description Civil Love Review Every so often you come across a movie you really enjoy. For me, this was it. This film contains everything a good movie needs: a good story, fine acting, direction, and a purpose. The theme of this film is that we are too quick to judge people sometimes and that people can occasionally surprise us. --Edwin L. Carpenter, Dove Foundation Movies & TV Genre for Featured Categories Drama0.0
    I want something emotional and heartfelt, but not boring. I don't want another slow, dull show.Surprised by Love A young businesswoman tries to convince her uptight parents to accept her boyfriend, but then she falls for an old high-school flame. Stars Hilarie Burton, Paul Campbell, and Tim Conway. IMDb 6.7 1 h 24 min 2015 7+ Drama Cheerful Passionate0.0
    I want something emotional and heartfelt, but not boring. I don't want another slow, dull show.All I Want for Christmas Tis the season for making dreams true. If everything goes according to plan for one young boy, he just may be sharing this Christmas with a new dad. Ever since her husband died, Sarah Armstrong has divided her time between working at a community center and caring for her nine-year-old son, Jesse. When Jesse wins the All I Want for Christmas contest sponsored by a toy company, his wish comes as no surprise to find a new dad. Just in time for the holidays, All I Want for Christmas is a surprise romantic treat for anyone who believes in the magic of love, friendship, and family. Movies & TV Holidays & Seasonal Christmas0.0
  • Loss: BinaryCrossEntropyLoss with these parameters:
    {
        "activation_fn": "torch.nn.modules.linear.Identity",
        "pos_weight": null
    }
    

Training Hyperparameters

Non-Default Hyperparameters

  • eval_strategy: epoch
  • per_device_train_batch_size: 32
  • learning_rate: 0.0001
  • weight_decay: 0.01
  • num_train_epochs: 1
  • lr_scheduler_type: cosine
  • warmup_ratio: 0.05
  • seed: 3407
  • bf16: True

All Hyperparameters

  • overwrite_output_dir: False
  • do_predict: False
  • eval_strategy: epoch
  • prediction_loss_only: True
  • per_device_train_batch_size: 32
  • per_device_eval_batch_size: 8
  • 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: 0.0001
  • 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: cosine
  • 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: 3407
  • data_seed: None
  • jit_mode_eval: False
  • use_ipex: 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: {'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
  • 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: False
  • 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: False
  • prompts: None
  • batch_sampler: batch_sampler
  • multi_dataset_batch_sampler: proportional
  • router_mapping: {}
  • learning_rate_mapping: {}

Training Logs

EpochStepTraining Lossamazon_test_ndcg@10amazon_val_ndcg@10
-1-1-0.6715-
0.0050501.8011--
0.01001000.8447--
0.01501500.6551--
0.01992000.6045--
0.02492500.5639--
0.02993000.5355--
0.03493500.5021--
0.03994000.5279--
0.04494500.4843--
0.04995000.4735--
0.05485500.4647--
0.05986000.4738--
0.06486500.4657--
0.06987000.4873--
0.07487500.4611--
0.07988000.439--
0.08478500.4302--
0.08979000.4417--
0.09479500.4461--
0.099710000.4277--
0.104710500.4367--
0.109711000.4273--
0.114711500.4104--
0.119612000.4259--
0.124612500.4412--
0.129613000.4075--
0.134613500.4307--
0.139614000.4473--
0.144614500.4502--
0.149615000.4168--
0.154515500.4174--
0.159516000.4229--
0.164516500.4276--
0.169517000.4152--
0.174517500.4236--
0.179518000.3981--
0.184418500.4082--
0.189419000.4261--
0.194419500.3888--
0.199420000.4314--
0.204420500.4254--
0.209421000.4296--
0.214421500.4324--
0.219322000.4074--
0.224322500.4127--
0.229323000.4206--
0.234323500.4078--
0.239324000.4015--
0.244324500.4316--
0.249325000.4067--
0.254225500.4066--
0.259226000.3945--
0.264226500.3927--
0.269227000.4264--
0.274227500.3822--
0.279228000.4176--
0.284128500.4128--
0.289129000.3914--
0.294129500.4091--
0.299130000.3949--
0.304130500.407--
0.309131000.3848--
0.314131500.4007--
0.319032000.4089--
0.324032500.3995--
0.329033000.3811--
0.334033500.3821--
0.339034000.3891--
0.344034500.3952--
0.349035000.4013--
0.353935500.4033--
0.358936000.3738--
0.363936500.3887--
0.368937000.3882--
0.373937500.3929--
0.378938000.3858--
0.383838500.3932--
0.388839000.3953--
0.393839500.393--
0.398840000.3937--
0.403840500.4102--
0.408841000.4216--
0.413841500.3825--
0.418742000.3997--
0.423742500.401--
0.428743000.3984--
0.433743500.4041--
0.438744000.388--
0.443744500.3947--
0.448745000.3988--
0.453645500.3826--
0.458646000.3981--
0.463646500.4018--
0.468647000.3696--
0.473647500.3739--
0.478648000.397--
0.483548500.4046--
0.488549000.3774--
0.493549500.3795--
0.498550000.3897--
0.503550500.3929--
0.508551000.3783--
0.513551500.389--
0.518452000.3832--
0.523452500.3477--
0.528453000.3694--
0.533453500.4011--
0.538454000.3978--
0.543454500.3776--
0.548455000.3939--
0.553355500.3656--
0.558356000.3815--
0.563356500.3664--
0.568357000.4057--
0.573357500.3859--
0.578358000.4032--
0.583358500.3853--
0.588259000.3679--
0.593259500.3748--
0.598260000.3839--
0.603260500.3719--
0.608261000.3626--
0.613261500.3991--
0.618162000.3927--
0.623162500.3923--
0.628163000.3649--
0.633163500.3843--
0.638164000.4072--
0.643164500.3875--
0.648165000.3903--
0.653065500.3658--
0.658066000.3889--
0.663066500.3823--
0.668067000.3789--
0.673067500.3623--
0.678068000.3865--
0.683068500.3809--
0.687969000.3833--
0.692969500.3752--
0.697970000.3834--
0.702970500.3667--
0.707971000.3791--
0.712971500.3689--
0.717872000.3663--
0.722872500.3995--
0.727873000.3696--
0.732873500.3747--
0.737874000.377--
0.742874500.3669--
0.747875000.3865--
0.752775500.3757--
0.757776000.3715--
0.762776500.3699--
0.767777000.363--
0.772777500.3861--
0.777778000.3518--
0.782778500.3801--
0.787679000.3881--
0.792679500.3797--
0.797680000.3788--
0.802680500.3761--
0.807681000.3838--
0.812681500.3664--
0.817582000.3745--
0.822582500.3706--
0.827583000.3771--
0.832583500.3865--
0.837584000.3601--
0.842584500.366--
0.847585000.3651--
0.852485500.366--
0.857486000.3676--
0.862486500.3828--
0.867487000.3374--
0.872487500.3644--
0.877488000.385--
0.882488500.3678--
0.887389000.3754--
0.892389500.3715--
0.897390000.3889--
0.902390500.367--
0.907391000.378--
0.912391500.3773--
0.917292000.388--
0.922292500.3549--
0.927293000.3726--
0.932293500.3787--
0.937294000.3863--
0.942294500.3705--
0.947295000.3643--
0.952195500.4003--
0.957196000.388--
0.962196500.3612--
0.967197000.3664--
0.972197500.3657--
0.977198000.3854--
0.982198500.3544--
0.987099000.3771--
0.992099500.3517--
0.9970100000.3659--
1.010030--0.9007
-1-1-0.9086-

Framework Versions

  • Python: 3.12.3
  • Sentence Transformers: 5.3.0
  • Transformers: 4.56.2
  • PyTorch: 2.11.0+cu128
  • Accelerate: 1.13.0
  • Datasets: 4.3.0
  • 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",
}

Как запустить

На своей GPU в ailist. Цена — по текущим предложениям провайдеров

GPU в ailist

Универсальный рантайм · sentence-transformers

Можно запустить

Подбираем GPU…

Класс архитектуры
XLMRobertaForSequenceClassification
Ревизия весов
0b87514

0 моделей