🩺 MediMaven BGE Cross-Encoder Reranker (v1.1)

A domain-adapted BGE reranker fine-tuned on synthetic triples (query, +, –) derived from 500 k medical passage pairs. This is a Cross Encoder model finetuned from BAAI/bge-reranker-base 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-base
Maximum Sequence Length: 512 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.


## 🚀 Usage

from sentence_transformers import CrossEncoder
model = CrossEncoder("medimaven-ai/medimaven-reranker-bge-cross-encoder")
scores = model.predict([("what causes gerd?", "Gastro-oesophageal reflux disease (GERD) occurs...")])
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

Training Recipe

Training Dataset

Medimaven-ai Dataset

Size: 570,914 training samples
Columns: sentence_0, sentence_1, and label

Approximate statistics based on the first 1000 samples:

	sentence_0	sentence_1	label
type	string	string	int
details	min: 17 characters mean: 98.9 characters max: 274 characters	min: 117 characters mean: 1647.18 characters max: 2339 characters	0: ~47.70% 1: ~52.30%

Samples:

sentence_0	sentence_1	label
`what is the life expectancy for babies with trisomy 18`	Most babies born with this condition die within the first few days or weeks of life, as they have so many medical complications. Just 5% to 10% make it past their first year. Like trisomy 18, no one knows why some babies get this condition. It's known that the chance increases with the mother's age, though women of any age can have a child with trisomy 13. About 80% of babies with trisomy 18 or 13 are born to mothers under 35. The condition can be diagnosed before birth with the same tests used to identify trisomy 18, or after birth by a physical examination. Trisomy 18 is a condition where you have three copies of each chromosome 18 in your body's cells instead of two. This can lead to serious physical and mental disabilities. There is no cure, though treatment can include surgeries, medicines, breathing tubes, and feeding tubes. Some parents opt just for comfort care. Life expectancy is usually a year or less. How old is the oldest living person with trisomy 18? The oldest people wer...	`1`
`what are some reasons doctors prescribe benzodiazepines`	But they can be habit-forming, especially if you take them regularly or for a long time. If you think you or a loved one may have a problem with benzodiazepine misuse, contact a doctor or a drug hotline. What is benzodiazepine abuse? Doctors define benzodiazepine abuse as using these drugs for non-medical reasons to get high, What is the incidence of benzodiazepine abuse? In a 12-month period between 2014 and 2015, more than 5 million people in the U.S. reported they had misused benzodiazepines. That's out of 30 million adults who used the drugs at all that year.	`0`

Loss: BinaryCrossEntropyLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 96
per_device_eval_batch_size: 96
num_train_epochs: 1

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: no
prediction_loss_only: True
per_device_train_batch_size: 96
per_device_eval_batch_size: 96
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: 1
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}
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: proportional

Training Logs

Epoch	Step	Training Loss
0.0841	500	0.5028
0.1681	1000	0.4081
0.2522	1500	0.3872
0.3362	2000	0.3738
0.4203	2500	0.3639
0.5044	3000	0.3551
0.5884	3500	0.3464
0.6725	4000	0.338
0.7566	4500	0.329
0.8406	5000	0.3297
0.9247	5500	0.3293

Framework Versions

Python: 3.11.12
Sentence Transformers: 4.1.0
Transformers: 4.52.2
PyTorch: 2.6.0+cu124
Accelerate: 1.7.0
Datasets: 2.14.4
Tokenizers: 0.21.1

📜 Citations

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",
}


@misc{medimaven-reranker-bge-cross-encoder,
  title        = {medimaven-reranker-bge-cross-encoder},
  author       = {Kyei-Mensah, Bernard},
  year         = {2025},
  howpublished = {\url{https://huggingface.co/dranreb1660/medimaven-reranker-bge-cross-encoder}}
}

dranreb1660
/

medimaven-reranker-bge-cross-encoder