---
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- feature-extraction
- sentence-similarity
- transformers
datasets:
- kornlu
language:
- ko
license: cc-by-4.0
---

# bi-matrix/gmatrix-embedding

해당 모델은 [KF-DeBERTa](https://huggingface.co/kakaobank/kf-deberta-base) 모델과 KorSTS, KorNLI 데이터셋을 활용하였으며, sentence-transformers의 공식 문서 내 소개된 [continue-learning](https://github.com/UKPLab/sentence-transformers/blob/master/examples/training/sts/training_stsbenchmark_continue_training.py) 방법을 통해 아래와 같이 학습되었습니다.
1. NLI 데이터셋을 통해 nagative sampling 후 MultipleNegativeRankingLoss 활용 및 STS 데이터셋을 통해 CosineSimilarityLoss를 활용하여 Multi-task Learning 학습 10epoch 진행
2. Learning Rate를 1e-06으로 줄여서 4epoch 추가 Multi-task 학습 진행

---
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.

<!--- Describe your model here -->

## Usage (Sentence-Transformers)

Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:

```
pip install -U sentence-transformers
```

Then you can use the model like this:

```python
from sentence_transformers import SentenceTransformer
sentences = ["This is an example sentence", "Each sentence is converted"]

model = SentenceTransformer("bi-matrix/gmatrix-embedding")
embeddings = model.encode(sentences)
print(embeddings)
```



## Usage (HuggingFace Transformers)
Without [sentence-transformers](https://www.SBERT.net), you can use the model like this: First, you pass your input through the transformer model, then you have to apply the right pooling-operation on-top of the contextualized word embeddings.

```python
from transformers import AutoTokenizer, AutoModel
import torch


#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0] #First element of model_output contains all token embeddings
    input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
    return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)


# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']

# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained("bi-matrix/gmatrix-embedding")
model = AutoModel.from_pretrained("bi-matrix/gmatrix-embedding")

# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')

# Compute token embeddings
with torch.no_grad():
    model_output = model(**encoded_input)

# Perform pooling. In this case, mean pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])

print("Sentence embeddings:")
print(sentence_embeddings)
```


## Evaluation Results

<!--- Describe how your model was evaluated -->

KorSTS 평가 데이터셋으로 평가한 결과입니다.

- Cosine Pearson: 85.77
- Cosine Spearman: 86.30
- Manhattan Pearson: 84.84
- Manhattan Spearman: 85.33
- Euclidean Pearson: 84.82
- Euclidean Spearman: 85.29
- Dot Pearson: 83.19
- Dot Spearman: 83.19

<br>

|model|cosine_pearson|cosine_spearman|euclidean_pearson|euclidean_spearman|manhattan_pearson|manhattan_spearman|dot_pearson|dot_spearman|
|:-------------------------|-----------------:|------------------:|--------------------:|---------------------:|--------------------:|---------------------:|--------------:|---------------:|
|[**gmatrix-embedding**](https://huggingface.co/bi-matrix/gmatrix-embedding)|**85.77**|**86.30**|**84.82**|**85.29**|**84.84**|**85.33**|**83.19**|**83.19**|
|[kf-deberta-multitask](https://huggingface.co/upskyy/kf-deberta-multitask)|85.75|86.25|84.79|85.25|84.80|85.27|82.93|82.86|
|[ko-sroberta-multitask](https://huggingface.co/jhgan/ko-sroberta-multitask)|84.77|85.6|83.71|84.40|83.70|84.38|82.42|82.33|
|[ko-sbert-multitask](https://huggingface.co/jhgan/ko-sbert-multitask)|84.13|84.71|82.42|82.66|82.41|82.69|80.05|79.69|
|[ko-sroberta-base-nli](https://huggingface.co/jhgan/ko-sroberta-nli)|82.83|83.85|82.87|83.29|82.88|83.28|80.34|79.69|
|[ko-sbert-nli](https://huggingface.co/jhgan/ko-sbert-multitask)|82.24|83.16|82.19|82.31|82.18|82.3|79.3|78.78|
|[ko-sroberta-sts](https://huggingface.co/jhgan/ko-sroberta-sts)|81.84|81.82|81.15|81.25|81.14|81.25|79.09|78.54|
|[ko-sbert-sts](https://huggingface.co/jhgan/ko-sbert-sts)|81.55|81.23|79.94|79.79|79.9|79.75|76.02|75.31|

<br>


<!--- Describe how your model was evaluated -->

G-MATRIX Embedding 데이터셋 측정 결과입니다.
사람 3명이서 0~5점으로 두 문장간의 유사도를 측정하여 점수를 내고 평균을 구하여 각 모델의 임베딩값을 통해

코사인 유사도, 유클리디안 거리, 맨하탄 거리, Dot-product를 구하여 피어슨, 스피어만 상관계수를 구한 값입니다.

- Cosine Pearson: 75.86
- Cosine Spearman: 65.75
- Manhattan Pearson: 72.65
- Manhattan Spearman: 65.20
- Euclidean Pearson: 72.48
- Euclidean Spearman: 65.32
- Dot Pearson: 64.71
- Dot Spearman: 53.90

<br>

model|cosine_pearson|cosine_spearman|euclidean_pearson|euclidean_spearman|manhattan_pearson|manhattan_spearman|dot_pearson|dot_spearman|
|:-------------------------|-----------------:|------------------:|--------------------:|---------------------:|--------------------:|---------------------:|--------------:|---------------:|
|[**gmatrix-embedding**](https://huggingface.co/bi-matrix/gmatrix-embedding)|**75.86**|**65.75**|**72.65**|**65.20**|**72.48**|**65.32**|**64.71**|**53.90**|
|[ko-sroberta-multitask](https://huggingface.co/jhgan/ko-sroberta-multitask)|71.78|63.16|70.80|63.47|70.89|63.72|53.57|44.23|
|[bge-m3](https://huggingface.co/BAAI/bge-m3)|64.15|60.65|61.88|60.68|61.88|60.19|64.16|60.71|

<br>



![image/png](https://cdn-uploads.huggingface.co/production/uploads/6350f6750b94548566da3279/CcK0QL3oQAz7sJOCtH6PB.png)

<br>

## G-MATRIX Embedding 레이블링 판단 기준 (KLUE-RoBERTa의 STS 데이터 생성 참고)
1. 두 문장의 유사한 정도를 보고 0~5점으로 판단
2. 맞춤법, 띄어쓰기, 온점이나 쉼표 차이는 판단 대상이 아님
3. 문장의 의도, 표현이 담고 있는 의미를 비교
4. 두 문장에 공통적으로 사용된 단어의 유무를 찾는 것이 아닌, 문장의 의미가 유사한지를 비교
5. 0은 의미적 유사성이 없는 경우이고, 5는 의미적으로 동등함을 뜻함



## Training
The model was trained with the parameters:

**DataLoader**:

`torch.utils.data.dataloader.DataLoader` of length 329 with parameters:
```
{'batch_size': 32, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}
```

**Loss**:

`sentence_transformers.losses.CosineSimilarityLoss.CosineSimilarityLoss` 


## Full Model Architecture
```
SentenceTransformer(
  (0): Transformer({'max_seq_length': 128, 'do_lower_case': True}) with Transformer model: DeBERTaV2Model 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False})
)
```

## Citing & Authors

<!--- Describe where people can find more information -->
[MINSANG SONG] at [BI-Matrix](https://www.bimatrix.co.kr/)