CrossEncoder based on nreimers/MiniLM-L3-H384-uncased

This is a Cross Encoder model finetuned from nreimers/MiniLM-L3-H384-uncased 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: nreimers/MiniLM-L3-H384-uncased
Maximum Sequence Length: 256 tokens
Number of Output Labels: 1 label

Model Sources

Documentation: Sentence Transformers Documentation
Documentation: Cross Encoder Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Cross Encoders on Hugging Face

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("lucasflins/CE-MiniLM-L3-H384-uncased-RankNet-20260213-175339")
# Get scores for pairs of texts
pairs = [
    ['maquina de lavar roupa panasonic', 'máquina de lavar 17kg f170b7wa branca 110v panasonic'],
    ['maquina de lavar roupa panasonic', 'máquina de lavar panasonic na-f180p7t 18kg'],
    ['maquina de lavar roupa panasonic', 'capa maquina lavar roupa impermeavel protetora panasonic 16kg naf160b6w zíper transparente preta'],
    ['maquina de lavar roupa panasonic', 'lavadora panasonic na-f180p7ta 18kg com tecnologia antibacteria cinza'],
    ['maquina de lavar roupa panasonic', 'lavadora de roupas panasonic f180p7t antibacteria ag 18kg'],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

# Or rank different texts based on similarity to a single text
ranks = model.rank(
    'maquina de lavar roupa panasonic',
    [
        'máquina de lavar 17kg f170b7wa branca 110v panasonic',
        'máquina de lavar panasonic na-f180p7t 18kg',
        'capa maquina lavar roupa impermeavel protetora panasonic 16kg naf160b6w zíper transparente preta',
        'lavadora panasonic na-f180p7ta 18kg com tecnologia antibacteria cinza',
        'lavadora de roupas panasonic f180p7t antibacteria ag 18kg',
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

Training Details

Training Dataset

Unnamed Dataset

Size: 58,401 training samples
Columns: query, text, and label

Approximate statistics based on the first 1000 samples:

	query	text	label
type	string	list	list
details	min: 2 characters mean: 23.86 characters max: 84 characters	min: 3 elements mean: 113.68 elements max: 647 elements	min: 3 elements mean: 113.68 elements max: 647 elements

Samples:

query	text	label
`mini berco e colchao`	`['moises mini berço malu para quarto do bebê com rodizios para meninos e meninas moderno com colchão multifuncional kids cor branco', 'mini berço multifuncional evolutivo 4 em 1 sem colchão bb040 completa móveis', 'mini berço moisés multifuncional mdf e rodízios com colchão', 'mini berço 2 gavetas com colchão moisés - branco', 'mini berço 100% mdf com colchão yescasa', ...]`	`[4.5160912172672525, 4.301410246007121, 4.149199637266272, 4.102236270603044, 4.0407219899682385, ...]`
`chinelo havaianas adulto original masculino`	`['chinelo havaianas masculino dual tiras larga ultra conforto', 'chinelo havaianas brasil adulto - branco', 'chinelo havaianas brasil branco adulto as legitimas original', 'chinelo de dedo adulto havaianas top basic fc', 'chinelo de dedo havaianas adulto dual masculino duas cores', ...]`	`[5.0, 1.3071868869897774, 1.2851885677229884, 1.2841688461219265, 1.262460870634281, ...]`
`armario emilly`	`['armário de cozinha completa 229cm rustic/cinza emilly madesa 017', 'armário de cozinha completa 229cm branco emilly madesa 17', 'armário de cozinha compacta 105cm com pia emilly madesa 04', 'armário de cozinha completa madesa emilly 293001 com balcão e paneleiro - rustic/preto', 'armário de cozinha completa com armário e balcão 229cm emilly madesa', ...]`	`[5.0, 4.8185494408696865, 4.732788739215357, 4.663854171188608, 4.617143051987073, ...]`

Loss: RankNetLoss with these parameters:

{
    "k": null,
    "sigma": 1.0,
    "eps": 1e-10,
    "reduction_log": "binary",
    "activation_fn": "torch.nn.modules.linear.Identity",
    "mini_batch_size": null
}

Evaluation Dataset

Unnamed Dataset

Size: 19,453 evaluation samples
Columns: query, text, and label

Approximate statistics based on the first 1000 samples:

	query	text	label
type	string	list	list
details	min: 3 characters mean: 23.46 characters max: 88 characters	min: 3 elements mean: 105.95 elements max: 374 elements	min: 3 elements mean: 105.95 elements max: 374 elements

Samples:

query	text	label
`maquina de lavar roupa panasonic`	`['máquina de lavar 17kg f170b7wa branca 110v panasonic', 'máquina de lavar panasonic na-f180p7t 18kg', 'capa maquina lavar roupa impermeavel protetora panasonic 16kg naf160b6w zíper transparente preta', 'lavadora panasonic na-f180p7ta 18kg com tecnologia antibacteria cinza', 'lavadora de roupas panasonic f180p7t antibacteria ag 18kg', ...]`	`[5.0, 0.7308349126503895, 0.6959629150629015, 0.6477526610672186, 0.5892796273761252, ...]`
`gyarda roupa`	`['guarda-roupa casal com espelho 3 portas de correr', 'guarda-roupa casal madesa reno 3 portas de correr com espelho', 'guarda roupa mongi 2 portas 2 gavetas branco caemmun', 'guarda-roupa casal com espelho 4 gavetas 2 portas de correr araplac benfica', 'guarda-roupa casal com espelho 6 portas', ...]`	`[5.0, 3.3783316133271213, 3.3299655932869037, 3.317338034957803, 3.1937724871830113, ...]`
`purificador de agua colormaq premium`	`['purificador de agua compressor premium narural e gelada colormaq', 'purificador de agua colormaq compressor branco 220v cpuhfba1', 'purificador de agua colormaq premium compressor cpuhfba1', 'purificador de água colormaq premium', 'purificador de água colormaq refrigerado por', ...]`	`[5.0, 2.7724961811106326, 2.3496201386993913, 1.9642276637675304, 1.154319875829175, ...]`

Loss: RankNetLoss with these parameters:

{
    "k": null,
    "sigma": 1.0,
    "eps": 1e-10,
    "reduction_log": "binary",
    "activation_fn": "torch.nn.modules.linear.Identity",
    "mini_batch_size": null
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_train_batch_size: 32
per_device_eval_batch_size: 16
num_train_epochs: 10
warmup_ratio: 0.1
log_level: info
tf32: True
load_best_model_at_end: True
hub_strategy: end
hub_private_repo: True
hub_always_push: True
eval_on_start: True

All Hyperparameters

Click to expand

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: 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
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 10
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: None
warmup_ratio: 0.1
warmup_steps: 0
log_level: info
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: False
fp16: False
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: False
fp16_full_eval: False
tf32: True
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: True
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
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: end
hub_private_repo: True
hub_always_push: True
hub_revision: None
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
include_for_metrics: []
eval_do_concat_batches: True
fp16_backend: auto
push_to_hub_model_id: None
push_to_hub_organization: None
mp_parameters:
auto_find_batch_size: False
full_determinism: False
torchdynamo: None
ray_scope: last
ddp_timeout: 1800
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
include_tokens_per_second: False
include_num_input_tokens_seen: no
neftune_noise_alpha: None
optim_target_modules: None
batch_eval_metrics: False
eval_on_start: True
use_liger_kernel: False
liger_kernel_config: None
eval_use_gather_object: False
average_tokens_across_devices: True
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss
0	0	-	1.0000
1.0	1826	0.9911	0.9819
2.0	3652	0.9698	0.9673
3.0	5478	0.9536	0.9577
4.0	7304	0.9418	0.9496
5.0	9130	0.9312	0.9455
6.0	10956	0.9235	0.9425
7.0	12782	0.9169	0.9416
8.0	14608	0.9125	0.9380
9.0	16434	0.9089	0.9377
10.0	18260	0.9074	0.9383

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.12.11
Sentence Transformers: 5.2.0
Transformers: 4.57.6
PyTorch: 2.7.1+cu128
Accelerate: 1.9.0
Datasets: 4.1.1
Tokenizers: 0.22.1

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

RankNetLoss

@inproceedings{burges2005learning,
  title={Learning to Rank using Gradient Descent},
  author={Burges, Chris and Shaked, Tal and Renshaw, Erin and Lazier, Ari and Deeds, Matt and Hamilton, Nicole and Hullender, Greg},
  booktitle={Proceedings of the 22nd international conference on Machine learning},
  pages={89--96},
  year={2005}
}

Downloads last month: 3

Safetensors

Model size

17.4M params

Tensor type

F32

Inference Providers NEW

Text Ranking

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for lucasflins/CE-MiniLM-L3-H384-uncased-RankNet-20260213-175339

Base model

nreimers/MiniLM-L3-H384-uncased

Finetuned

(4)

this model

Paper for lucasflins/CE-MiniLM-L3-H384-uncased-RankNet-20260213-175339

Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks

Paper • 1908.10084 • Published Aug 27, 2019 • 14