---
tags:
- sentence-transformers
- sentence-similarity
- feature-extraction
- dense
- generated_from_trainer
- dataset_size:19380
- loss:MultipleNegativesRankingLoss
base_model: intfloat/multilingual-e5-large
widget:
- source_sentence: 'query: ASC X.12 는 뭔가요?'
sentences:
- 'passage: Accredited Standard Committee X.12'
- 'passage: BCP Measurement and statistics Handling'
- 'passage: Bearer Inter Working Function'
- source_sentence: 'query: BECN 뜻 설명해줘.'
sentences:
- 'passage: AU Physical Control Block'
- 'passage: Backward Explicit Congestion Notification'
- 'passage: Beginning-Of-Tape Marker'
- source_sentence: 'query: BMD 뜻 설명해줘.'
sentences:
- 'passage: 5th Generation Computer'
- 'passage: Billing Mediation Device'
- 'passage: 3 Dimensional Television'
- source_sentence: 'query: 5GL 는 뭔가요?'
sentences:
- 'passage: Antenna Front-end Combiner Unit'
- 'passage: Authentication Center'
- 'passage: 5th Generation programming Language'
- source_sentence: 'query: 무슨 뜻이야 BCHB?'
sentences:
- 'passage: Assisted-Global Navigation Satellite System'
- 'passage: BCP Configuration Handler Block'
- 'passage: ATM Link Processor'
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
- cosine_accuracy@1
- cosine_accuracy@3
- cosine_accuracy@5
- cosine_accuracy@10
- cosine_precision@1
- cosine_precision@3
- cosine_precision@5
- cosine_precision@10
- cosine_recall@1
- cosine_recall@3
- cosine_recall@5
- cosine_recall@10
- cosine_ndcg@10
- cosine_mrr@10
- cosine_map@100
model-index:
- name: SentenceTransformer based on intfloat/multilingual-e5-large
results:
- task:
type: information-retrieval
name: Information Retrieval
dataset:
name: e5 eval real
type: e5-eval-real
metrics:
- type: cosine_accuracy@1
value: 0.8415
name: Cosine Accuracy@1
- type: cosine_accuracy@3
value: 0.9715
name: Cosine Accuracy@3
- type: cosine_accuracy@5
value: 0.985
name: Cosine Accuracy@5
- type: cosine_accuracy@10
value: 0.994
name: Cosine Accuracy@10
- type: cosine_precision@1
value: 0.8415
name: Cosine Precision@1
- type: cosine_precision@3
value: 0.3238333333333333
name: Cosine Precision@3
- type: cosine_precision@5
value: 0.19700000000000004
name: Cosine Precision@5
- type: cosine_precision@10
value: 0.09940000000000002
name: Cosine Precision@10
- type: cosine_recall@1
value: 0.8415
name: Cosine Recall@1
- type: cosine_recall@3
value: 0.9715
name: Cosine Recall@3
- type: cosine_recall@5
value: 0.985
name: Cosine Recall@5
- type: cosine_recall@10
value: 0.994
name: Cosine Recall@10
- type: cosine_ndcg@10
value: 0.9288608308614111
name: Cosine Ndcg@10
- type: cosine_mrr@10
value: 0.9068103174603175
name: Cosine Mrr@10
- type: cosine_map@100
value: 0.9070514070699416
name: Cosine Map@100
---
# SentenceTransformer based on intfloat/multilingual-e5-large
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [intfloat/multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large) on the train dataset. It maps sentences & paragraphs to a 1024-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:** [intfloat/multilingual-e5-large](https://huggingface.co/intfloat/multilingual-e5-large)
- **Maximum Sequence Length:** 256 tokens
- **Output Dimensionality:** 1024 dimensions
- **Similarity Function:** Cosine Similarity
- **Training Dataset:**
- train
### Model Sources
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
### Full Model Architecture
```
SentenceTransformer(
(0): Transformer({'max_seq_length': 256, 'do_lower_case': False, 'architecture': 'XLMRobertaModel'})
(1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
(2): Normalize()
)
```
## 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 SentenceTransformer
# Download from the 🤗 Hub
model = SentenceTransformer("sentence_transformers_model_id")
# Run inference
sentences = [
'query: 무슨 뜻이야 BCHB?',
'passage: BCP Configuration Handler Block',
'passage: Assisted-Global Navigation Satellite System',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.8122, 0.0080],
# [0.8122, 1.0000, 0.0858],
# [0.0080, 0.0858, 1.0000]])
```
## Evaluation
### Metrics
#### Information Retrieval
* Dataset: `e5-eval-real`
* Evaluated with [InformationRetrievalEvaluator](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.InformationRetrievalEvaluator)
| Metric | Value |
|:--------------------|:-----------|
| cosine_accuracy@1 | 0.8415 |
| cosine_accuracy@3 | 0.9715 |
| cosine_accuracy@5 | 0.985 |
| cosine_accuracy@10 | 0.994 |
| cosine_precision@1 | 0.8415 |
| cosine_precision@3 | 0.3238 |
| cosine_precision@5 | 0.197 |
| cosine_precision@10 | 0.0994 |
| cosine_recall@1 | 0.8415 |
| cosine_recall@3 | 0.9715 |
| cosine_recall@5 | 0.985 |
| cosine_recall@10 | 0.994 |
| **cosine_ndcg@10** | **0.9289** |
| cosine_mrr@10 | 0.9068 |
| cosine_map@100 | 0.9071 |
## Training Details
### Training Dataset
#### train
* Dataset: train
* Size: 19,380 training samples
* Columns: 0 and 1
* Approximate statistics based on the first 1000 samples:
| | 0 | 1 |
|:--------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|
| type | string | string |
| details |
query: (e)CSFB 알려줘 | passage: (enhanced) Circuit Switched Fallback |
| query: 1000 BASE 알려줘 | passage: 1000 Base Standard |
| query: 1080i 뜻 설명해줘. | passage: 1080 interlace scan |
* Loss: [MultipleNegativesRankingLoss](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
```json
{
"scale": 20.0,
"similarity_fct": "cos_sim",
"gather_across_devices": false
}
```
### Training Hyperparameters
#### Non-Default Hyperparameters
- `eval_strategy`: steps
- `per_device_train_batch_size`: 64
- `per_device_eval_batch_size`: 64
- `learning_rate`: 1e-05
- `weight_decay`: 0.01
- `lr_scheduler_type`: cosine
- `warmup_ratio`: 0.1
- `bf16`: True
- `batch_sampler`: no_duplicates
#### All Hyperparameters