CrossEncoder

This is a Cross Encoder model trained 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
Maximum Sequence Length: 8192 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("cross_encoder_model_id")
# Get scores for pairs of texts
pairs = [
    ['Hexacarboxylporphyrin/Creatinine [Molar ratio] in 24 hour Urine', '[Molar ratio] in Hexacarboxylporphyrin/Creatinine 24 hour Urn'],
    ['HLA-A2 Ql (Bld/Tiss donor)', 'HLA-A11 donor) (Bld/Tiss Ql'],
    ['Urea nitrogen [Mass/volume] in Urine', 'POC Urine Urea nitrogen Measurement'],
    ['Cauliflower IgG (S) [Mass/Vol]', 'POC Cauliflower Immune globulin (S) Radioallergosorbent'],
    ['Mannose-binding protein [Mass/volume] in Serum', 'MBP Level Serum Quantitative'],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

# Or rank different texts based on similarity to a single text
ranks = model.rank(
    'Hexacarboxylporphyrin/Creatinine [Molar ratio] in 24 hour Urine',
    [
        '[Molar ratio] in Hexacarboxylporphyrin/Creatinine 24 hour Urn',
        'HLA-A11 donor) (Bld/Tiss Ql',
        'POC Urine Urea nitrogen Measurement',
        'POC Cauliflower Immune globulin (S) Radioallergosorbent',
        'MBP Level Serum Quantitative',
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

Training Details

Training Dataset

Unnamed Dataset

Size: 1,120,911 training samples
Columns: anchor, positive, and label

Approximate statistics based on the first 1000 samples:

	anchor	positive	label
type	string	string	int
details	min: 7 characters mean: 50.32 characters max: 140 characters	min: 8 characters mean: 44.25 characters max: 119 characters	0: ~42.50% 1: ~57.50%

Samples:

anchor	positive	label
`cloZAPine cutoff Confirm (U) [Mass/Vol]`	`Clozaril cutoff (U) Confirmation [Mass/Vol]`	`1`
`Horse hair+Horse dander IgE RAST class (S)`	`Serum Horse hair+Horse dander IgE Ab`	`0`
`Deprecated Red Kidney Bean IgG Ab RAST class [Presence] in Serum`	`Red Kidney Bean IgG Ab [Presence] Radioallergosorbent test`	`0`

Loss: BinaryCrossEntropyLoss with these parameters:

{
    "activation_fn": "torch.nn.modules.linear.Identity",
    "pos_weight": 5
}

Evaluation Dataset

Unnamed Dataset

Size: 100,000 evaluation samples
Columns: anchor, positive, and label

Approximate statistics based on the first 1000 samples:

	anchor	positive	label
type	string	string	int
details	min: 12 characters mean: 47.94 characters max: 133 characters	min: 7 characters mean: 42.78 characters max: 124 characters	0: ~43.70% 1: ~56.30%

Samples:

anchor	positive	label
`Hexacarboxylporphyrin/Creatinine [Molar ratio] in 24 hour Urine`	`[Molar ratio] in Hexacarboxylporphyrin/Creatinine 24 hour Urn`	`1`
`HLA-A2 Ql (Bld/Tiss donor)`	`HLA-A11 donor) (Bld/Tiss Ql`	`0`
`Urea nitrogen [Mass/volume] in Urine`	`POC Urine Urea nitrogen Measurement`	`0`

Loss: BinaryCrossEntropyLoss with these parameters:

{
    "activation_fn": "torch.nn.modules.linear.Identity",
    "pos_weight": 5
}

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 64
num_train_epochs: 1
learning_rate: 1e-07
warmup_steps: 0.1
bf16: True
eval_strategy: steps
per_device_eval_batch_size: 64
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

per_device_train_batch_size: 64
num_train_epochs: 1
max_steps: -1
learning_rate: 1e-07
lr_scheduler_type: linear
lr_scheduler_kwargs: None
warmup_steps: 0.1
optim: adamw_torch_fused
optim_args: None
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
optim_target_modules: None
gradient_accumulation_steps: 1
average_tokens_across_devices: True
max_grad_norm: 1.0
label_smoothing_factor: 0.0
bf16: True
fp16: False
bf16_full_eval: False
fp16_full_eval: False
tf32: None
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
use_liger_kernel: False
liger_kernel_config: None
use_cache: False
neftune_noise_alpha: None
torch_empty_cache_steps: None
auto_find_batch_size: False
log_on_each_node: True
logging_nan_inf_filter: True
include_num_input_tokens_seen: no
log_level: passive
log_level_replica: warning
disable_tqdm: False
project: huggingface
trackio_space_id: trackio
eval_strategy: steps
per_device_eval_batch_size: 64
prediction_loss_only: True
eval_on_start: False
eval_do_concat_batches: True
eval_use_gather_object: False
eval_accumulation_steps: None
include_for_metrics: []
batch_eval_metrics: False
save_only_model: False
save_on_each_node: False
enable_jit_checkpoint: False
push_to_hub: False
hub_private_repo: None
hub_model_id: None
hub_strategy: every_save
hub_always_push: False
hub_revision: None
load_best_model_at_end: False
ignore_data_skip: False
restore_callback_states_from_checkpoint: False
full_determinism: False
seed: 42
data_seed: None
use_cpu: False
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
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_pin_memory: True
dataloader_persistent_workers: False
dataloader_prefetch_factor: None
remove_unused_columns: True
label_names: None
train_sampling_strategy: random
length_column_name: length
ddp_find_unused_parameters: None
ddp_bucket_cap_mb: None
ddp_broadcast_buffers: False
ddp_backend: None
ddp_timeout: 1800
fsdp: []
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
deepspeed: None
debug: []
skip_memory_metrics: True
do_predict: False
resume_from_checkpoint: None
warmup_ratio: None
local_rank: -1
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss
0.1000	1751	1.5995	1.2376
0.1999	3502	1.1824	1.0858
0.2999	5253	1.0610	1.0036
0.3999	7004	1.0037	0.9503
0.4999	8755	0.9602	0.9021
0.5998	10506	0.9261	0.8669
0.6998	12257	0.8943	0.8422
0.7998	14008	0.8777	0.8264
0.8997	15759	0.8619	0.8176
0.9997	17510	0.8668	0.8150
1.0	17515	-	0.8150

Framework Versions

Python: 3.10.20
Sentence Transformers: 5.3.0
Transformers: 5.4.0
PyTorch: 2.10.0+cu128
Accelerate: 1.13.0
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",
}

Downloads last month: 1,128

Safetensors

Model size

0.6B params

Tensor type

F32

Inference Providers NEW

Text Ranking

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

Paper for NCHS/ttc-reranker-mvp

Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks

Paper • 1908.10084 • Published Aug 27, 2019 • 12