---
tags:
- sentence-transformers
- cross-encoder
- reranker
- generated_from_trainer
- dataset_size:1047
- loss:MultipleNegativesRankingLoss
base_model: Qwen/Qwen3-Embedding-0.6B
pipeline_tag: text-ranking
library_name: sentence-transformers
metrics:
- accuracy
- accuracy_threshold
- f1
- f1_threshold
- precision
- recall
- average_precision
model-index:
- name: CrossEncoder based on Qwen/Qwen3-Embedding-0.6B
results:
- task:
type: cross-encoder-classification
name: Cross Encoder Classification
dataset:
name: Unknown
type: unknown
metrics:
- type: accuracy
value: 0.9389312977099237
name: Accuracy
- type: accuracy_threshold
value: 0.726271390914917
name: Accuracy Threshold
- type: f1
value: 0.9391634980988592
name: F1
- type: f1_threshold
value: 0.726271390914917
name: F1 Threshold
- type: precision
value: 0.9356060606060606
name: Precision
- type: recall
value: 0.9427480916030534
name: Recall
- type: average_precision
value: 0.9508539647615596
name: Average Precision
- type: accuracy
value: 0.9435975609756098
name: Accuracy
- type: accuracy_threshold
value: 0.8168901205062866
name: Accuracy Threshold
- type: f1
value: 0.944693572496263
name: F1
- type: f1_threshold
value: 0.7354934215545654
name: F1 Threshold
- type: precision
value: 0.9266862170087976
name: Precision
- type: recall
value: 0.9634146341463414
name: Recall
- type: average_precision
value: 0.9544295903264528
name: Average Precision
---
# CrossEncoder based on Qwen/Qwen3-Embedding-0.6B
This is a [Cross Encoder](https://www.sbert.net/docs/cross_encoder/usage/usage.html) model finetuned from [Qwen/Qwen3-Embedding-0.6B](https://huggingface.co/Qwen/Qwen3-Embedding-0.6B) using the [sentence-transformers](https://www.SBERT.net) 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:** [Qwen/Qwen3-Embedding-0.6B](https://huggingface.co/Qwen/Qwen3-Embedding-0.6B)
- **Maximum Sequence Length:** 32768 tokens
- **Number of Output Labels:** 1 label
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Documentation:** [Cross Encoder Documentation](https://www.sbert.net/docs/cross_encoder/usage/usage.html)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/huggingface/sentence-transformers)
- **Hugging Face:** [Cross Encoders on Hugging Face](https://huggingface.co/models?library=sentence-transformers&other=cross-encoder)
## Usage
### Direct Usage (Sentence Transformers)
First install the Sentence Transformers library:
```bash
pip install -U sentence-transformers
```
Then you can load this model and run inference.
```python
from sentence_transformers import CrossEncoder
# Download from the π€ Hub
model = CrossEncoder("vkimbris/qwen3_06b_items_reranker")
# Get scores for pairs of texts
pairs = [
['ΠΠ°ΡΠ°Π±ΠΈ ΠΏΠΎΡΠΎΡΠΎΠΊ Π³ΠΎΡΡΠΈΡΠ½ΡΠΉ ΠΡΠ΅ΠΌΠΈΡΠΌ Fumiko Resfood 1ΠΊΠ³, 10ΡΡ/ΠΊΠΎΡ, ΠΠΈΡ
Π°ΠΉ', 'ΠΠ°ΡΠ°Π±ΠΈ Fumiko Premium Π³ΡΠ΅ΠΉΠ΄ Π, 85% Ρ
ΡΠ΅Π½Π°'],
['Π‘ΠΎΡΡ Π’Π΅ΡΠΈΡΠΊΠΈ Genso 1,5n/1,7ΠΊΠ³, Π±ΡΡ/ΠΊΠΎΡ, Π ΠΎΡΡΠΈΡ', 'Π‘ΠΎΡΡ Π’Π΅ΡΠΈΡΠΊΠΈ Genso'],
['Π£ΠΊΡΡΡ ΡΠΈΡΠΎΠ²ΡΠΉ Padam Prem Resfood 20Π», Π ΠΎΡΡΠΈΡ', 'Π£ΠΊΡΡΡ ΡΠΈΡΠΎΠ²ΡΠΉ Padam Premium'],
['ΠΠΌΠ±ΠΈΡΡ ΠΌΠ°ΡΠΈΠ½ΠΎΠ²Π°Π½Π½ΡΠΉ ΡΠΎΠ·ΠΎΠ²ΡΠΉ Tabuko Restood 1,5 ΠΊΠ³, Π²Π΅Ρ ΡΡΡ
ΠΎΠ³ΠΎ Π²Π΅Ρ-Π²Π° 1ΠΊΠ³, 10ΡΡ/ΠΊΠΎΡ, ΠΠΈΡΠ°ΠΉ', 'ΠΠΌΠ±ΠΈΡΡ ΠΌΠ°ΡΠΈΠ½ΠΎΠ²Π°Π½Π½ΡΠΉ Tabuko ΡΠΎΠ·ΠΎΠ²ΡΠΉ'],
["ΠΠ°ΡΡΠ° Π’ΠΎΠΌ Π―ΠΌ 'Genso' ΠΏΠ°ΠΊΠ΅Ρ (0,400 ΠΊΠ³) ΡΠΏΠ°ΠΊ. 24 ΡΡ. Π’Π°ΠΉΠ»Π°Π½Π΄", 'ΠΠ°ΡΡΠ° Π’ΠΎΠΌ Π―ΠΌ Genso'],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)
# Or rank different texts based on similarity to a single text
ranks = model.rank(
'ΠΠ°ΡΠ°Π±ΠΈ ΠΏΠΎΡΠΎΡΠΎΠΊ Π³ΠΎΡΡΠΈΡΠ½ΡΠΉ ΠΡΠ΅ΠΌΠΈΡΠΌ Fumiko Resfood 1ΠΊΠ³, 10ΡΡ/ΠΊΠΎΡ, ΠΠΈΡ
Π°ΠΉ',
[
'ΠΠ°ΡΠ°Π±ΠΈ Fumiko Premium Π³ΡΠ΅ΠΉΠ΄ Π, 85% Ρ
ΡΠ΅Π½Π°',
'Π‘ΠΎΡΡ Π’Π΅ΡΠΈΡΠΊΠΈ Genso',
'Π£ΠΊΡΡΡ ΡΠΈΡΠΎΠ²ΡΠΉ Padam Premium',
'ΠΠΌΠ±ΠΈΡΡ ΠΌΠ°ΡΠΈΠ½ΠΎΠ²Π°Π½Π½ΡΠΉ Tabuko ΡΠΎΠ·ΠΎΠ²ΡΠΉ',
'ΠΠ°ΡΡΠ° Π’ΠΎΠΌ Π―ΠΌ Genso',
]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]
```
## Evaluation
### Metrics
#### Cross Encoder Classification
* Evaluated with [CrossEncoderClassificationEvaluator](https://sbert.net/docs/package_reference/cross_encoder/evaluation.html#sentence_transformers.cross_encoder.evaluation.CrossEncoderClassificationEvaluator)
| Metric | Value |
|:----------------------|:-----------|
| accuracy | 0.9389 |
| accuracy_threshold | 0.7263 |
| f1 | 0.9392 |
| f1_threshold | 0.7263 |
| precision | 0.9356 |
| recall | 0.9427 |
| **average_precision** | **0.9509** |
#### Cross Encoder Classification
* Evaluated with [CrossEncoderClassificationEvaluator](https://sbert.net/docs/package_reference/cross_encoder/evaluation.html#sentence_transformers.cross_encoder.evaluation.CrossEncoderClassificationEvaluator)
| Metric | Value |
|:----------------------|:-----------|
| accuracy | 0.9436 |
| accuracy_threshold | 0.8169 |
| f1 | 0.9447 |
| f1_threshold | 0.7355 |
| precision | 0.9267 |
| recall | 0.9634 |
| **average_precision** | **0.9544** |
## Training Details
### Training Dataset
#### Unnamed Dataset
* Size: 1,047 training samples
* Columns: premise and hypothesis
* Approximate statistics based on the first 1000 samples:
| | premise | hypothesis |
|:--------|:------------------------------------------------------------------------------------------------|:----------------------------------------------------------------------------------------------|
| type | string | string |
| details |
Π‘ΠΌΠ΅ΡΡ ΠΌΡΡΠ½Π°Ρ ΡΠ΅ΠΌΠΏΡΡΠ½Π°Ρ 'KANESHIRO' 1ΠΊΠ³ | ΠΡΠΊΠ° ΡΠ΅ΠΌΠΏΡΡΠ½Π°Ρ Kaneshiro |
| Π‘ΠΌΠ΅ΡΡ ΡΠ΅ΠΌΠΏΡΡΠ½Π°Ρ Kaneshiro Resfood 1xr. 10ΡΡ/ΠΊΠΎΡ | ΠΡΠΊΠ° ΡΠ΅ΠΌΠΏΡΡΠ½Π°Ρ Kaneshiro |
| ΠΠΌΠ±ΠΈΡΡ ΠΌΠ°ΡΠΈΠ½ΠΎΠ²Π°Π½Π½ΡΠΉ ΡΠΎΠ·ΠΎΠ²ΡΠΉ 'Hansey' 1,4 ΠΊΠ³*10 (Π².Ρ. ΠΠΠ ΠΠΠΠ ΠΠ 10 ΠΠΠ§ΠΠ) | ΠΠΌΠ±ΠΈΡΡ ΠΌΠ°ΡΠΈΠ½ΠΎΠ²Π°Π½Π½ΡΠΉ ΡΠΎΠ·ΠΎΠ²ΡΠΉ Hansey, Π²Π΅Ρ ΡΡΡ
ΠΎΠ³ΠΎ Π²Π΅ΡΠ΅ΡΡΠ²Π° 1000 Π³ |
* Loss: [MultipleNegativesRankingLoss](https://sbert.net/docs/package_reference/cross_encoder/losses.html#multiplenegativesrankingloss) with these parameters:
```json
{
"scale": 10.0,
"num_negatives": 4,
"activation_fn": "torch.nn.modules.activation.Sigmoid"
}
```
### Evaluation Dataset
#### Unnamed Dataset
* Size: 262 evaluation samples
* Columns: premise and hypothesis
* Approximate statistics based on the first 262 samples:
| | premise | hypothesis |
|:--------|:------------------------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------------------|
| type | string | string |
| details | ΠΠ°ΡΠ°Π±ΠΈ ΠΏΠΎΡΠΎΡΠΎΠΊ Π³ΠΎΡΡΠΈΡΠ½ΡΠΉ ΠΡΠ΅ΠΌΠΈΡΠΌ Fumiko Resfood 1ΠΊΠ³, 10ΡΡ/ΠΊΠΎΡ, ΠΠΈΡ
Π°ΠΉ | ΠΠ°ΡΠ°Π±ΠΈ Fumiko Premium Π³ΡΠ΅ΠΉΠ΄ Π, 85% Ρ
ΡΠ΅Π½Π° |
| Π‘ΠΎΡΡ Π’Π΅ΡΠΈΡΠΊΠΈ Genso 1,5n/1,7ΠΊΠ³, Π±ΡΡ/ΠΊΠΎΡ, Π ΠΎΡΡΠΈΡ | Π‘ΠΎΡΡ Π’Π΅ΡΠΈΡΠΊΠΈ Genso |
| Π£ΠΊΡΡΡ ΡΠΈΡΠΎΠ²ΡΠΉ Padam Prem Resfood 20Π», Π ΠΎΡΡΠΈΡ | Π£ΠΊΡΡΡ ΡΠΈΡΠΎΠ²ΡΠΉ Padam Premium |
* Loss: [MultipleNegativesRankingLoss](https://sbert.net/docs/package_reference/cross_encoder/losses.html#multiplenegativesrankingloss) with these parameters:
```json
{
"scale": 10.0,
"num_negatives": 4,
"activation_fn": "torch.nn.modules.activation.Sigmoid"
}
```
### Training Hyperparameters
#### Non-Default Hyperparameters
- `eval_strategy`: steps
- `per_device_train_batch_size`: 16
- `per_device_eval_batch_size`: 16
- `learning_rate`: 2e-05
- `num_train_epochs`: 15
- `warmup_ratio`: 0.1
- `fp16`: True
#### All Hyperparameters