SentenceTransformer based on Qwen/Qwen3-Embedding-4B

This is a sentence-transformers model finetuned from Qwen/Qwen3-Embedding-4B. It maps sentences & paragraphs to a 2560-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: Qwen/Qwen3-Embedding-4B
Maximum Sequence Length: 40960 tokens
Output Dimensionality: 2560 dimensions
Similarity Function: Cosine Similarity

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': 40960, 'do_lower_case': False, 'architecture': 'Qwen3Model'})
  (1): Pooling({'word_embedding_dimension': 2560, '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("novacardsai/qwen3-med-classifier")
# Run inference
queries = [
    "Instruct: Classify this medical flashcard into one or more relevant categories.\nQuery: picture frame vertebra in mixed phase of pagets disease",
]
documents = [
    'Instruct: Classify this medical flashcard into one or more relevant categories.\nQuery: do patients with radial head subluxation experience sensory loss over the dorsal side of the hand or wrist? no.',
    'Instruct: Classify this medical flashcard into one or more relevant categories.\nQuery: bronchiolitis obliterans (constrictive) is a patchy chronic inflammation & fibrosis of the bronchioles, which leads to collapse/obliteration of the bronchioles',
    'Instruct: Classify this medical flashcard into one or more relevant categories.\nQuery: diseases such as minimal change disease, focal segmental glomerulosclerosis, and membranous nephropathy fall under the category of podocytopathies, which affect the podocytes of the kidney.',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 2560] [3, 2560]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)
# tensor([[0.9688, 0.5820, 0.5742]], dtype=torch.bfloat16)

Evaluation

Metrics

Triplet

Dataset: med_flashcard_val
Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.962

Training Details

Training Dataset

Unnamed Dataset

Size: 65,724 training samples
Columns: sentence_0, sentence_1, and sentence_2

Approximate statistics based on the first 1000 samples:

	sentence_0	sentence_1	sentence_2
type	string	string	string
details	min: 21 tokens mean: 43.47 tokens max: 203 tokens	min: 22 tokens mean: 42.65 tokens max: 238 tokens	min: 21 tokens mean: 43.69 tokens max: 268 tokens

Samples:

sentence_0	sentence_1	sentence_2
`Instruct: Classify this medical flashcard into one or more relevant categories. Query: tobacco juice color amniotic fluid is seen in intrauterine death of fetus`	`Instruct: Classify this medical flashcard into one or more relevant categories. Query: in cases of acute fatty liver of pregnancy, an increase in total bilirubin levels is commonly observed.`	`Instruct: Classify this medical flashcard into one or more relevant categories. Query: during pregnancy, avoid these vaccines: measles, mumps, and rubella (mmr), varicella, polio, and intranasal influenza.`
`Instruct: Classify this medical flashcard into one or more relevant categories. Query: 10 no. karman cannula is used at 11 weeks of pregnancy`	`Instruct: Classify this medical flashcard into one or more relevant categories. Query: a persistent moro reflex is observed in individuals with cerebral palsy.`	`Instruct: Classify this medical flashcard into one or more relevant categories. Query: what enzyme in macrophages is primarily used to destroy phagocytosed material? lysozyme`
`Instruct: Classify this medical flashcard into one or more relevant categories. Query: the presence of the dc-sign receptor increases susceptibility to hiv infection.`	`Instruct: Classify this medical flashcard into one or more relevant categories. Query: what is the name of the blood transfusion reaction that typically manifests with symptoms including fever, headaches, chills, and flushing? it is called a febrile nonhemolytic transfusion reaction.`	`Instruct: Classify this medical flashcard into one or more relevant categories. Query: which medications are used in the treatment of acute bipolar depression? 1. lurasidone2. olanzapine + fluoxetine3. lithium4. valproate5. quetiapine6. lamotrigine`

Loss: MultipleNegativesRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
num_train_epochs: 1
multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 8
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: 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
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}
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
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: 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: False
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: round_robin
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	med_flashcard_val_cosine_accuracy
0.0609	500	1.9105	-
0.1217	1000	1.4749	-
0.1826	1500	1.3069	-
0.2000	1643	-	0.9482
0.2434	2000	1.202	-
0.3043	2500	1.1576	-
0.3651	3000	1.1167	-
0.4000	3286	-	0.9544
0.4260	3500	1.0333	-
0.4869	4000	1.0287	-
0.5477	4500	1.0063	-
0.5999	4929	-	0.9620
0.6086	5000	0.9879	-
0.6694	5500	0.9673	-
0.7303	6000	0.953	-
0.7911	6500	0.9558	-
0.7999	6572	-	0.9626
0.8520	7000	0.9696	-
0.9129	7500	0.9525	-
0.9737	8000	0.9641	-
0.9999	8215	-	0.9620
1.0	8216	-	0.9620

Framework Versions

Python: 3.12.12
Sentence Transformers: 5.1.2
Transformers: 4.57.3
PyTorch: 2.9.0+cu126
Accelerate: 1.12.0
Datasets: 4.0.0
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",
}

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}
}