Gte Base Ko
A sentence embedding model fine-tuned on the Korean triplet dataset based on the Alibaba-NLP/gte-multilingual-base model for semantic similarity calculation
Text Embedding
Safetensors Supports Multiple Languages#Korean sentence embedding#Semantic similarity calculation#Multilingual support
Downloads 18
Release Time : 11/17/2024
Model Overview
This is a sentence transformer model fine-tuned on the Korean triplet dataset nlpai-lab/ko-triplet-v1.0 based on the Alibaba-NLP/gte-multilingual-base model. It maps sentences and paragraphs to a 768-dimensional dense vector space and can be used for tasks such as semantic text similarity, semantic search, and text classification.
Model Features
Korean optimization
Optimized specifically for Korean text and fine-tuned on the Korean triplet dataset
Long text support
Supports a sequence length of up to 8192 tokens, suitable for processing long texts
High accuracy
Achieved a cosine accuracy of 98.55% on the evaluation dataset
Model Capabilities
Semantic text similarity calculation
Semantic search
Text classification
Cluster analysis
Feature extraction
Use Cases
Information retrieval
Similar document retrieval
Find semantically similar documents based on the query text
Text analysis
Text clustering
Automatically group semantically similar texts
๐ SentenceTransformer based on Alibaba-NLP/gte-multilingual-base
This is a sentence-transformers model fine - tuned from Alibaba-NLP/gte-multilingual-base on the nlpai-lab/ko-triplet-v1.0 dataset. It maps sentences and paragraphs to a 768 - dimensional dense vector space, which can be applied to semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
๐ Quick Start
First, you need to 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 ๐ค Hub
model = SentenceTransformer("scottsuk0306/gte-base-ko")
# Run inference
sentences = [
'์ค, ์ธ์ฌํ๋ค, ์์ฃผํ๋ค, ์น๋ค, ์ฌ๋, ๋ชฉํ',
'๊ทธ ์ค์ ์์ฃผํ ์ฌ๋์๊ฒ ๋ชฉํ์ ์น๋ฉฐ ์ธ์ฌ๋ฅผ ํ๋ค.',
'์ฌ์ฃผ๋ผ๋ ๊ฒ์ ์์ , ์์
, ํ์๊ธฐ, ๊ธ์ฐ๊ธฐ, ๋งํ๊ธฐ, ๊ณ์ฐํ๊ธฐ๋ค.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
โจ Features
- Maps sentences and paragraphs to a 768 - dimensional dense vector space.
- Can be used for multiple NLP tasks such as semantic textual similarity, semantic search, etc.
๐ฆ Installation
pip install -U sentence-transformers
๐ป Usage Examples
Basic Usage
from sentence_transformers import SentenceTransformer
# Download from the ๐ค Hub
model = SentenceTransformer("scottsuk0306/gte-base-ko")
# Run inference
sentences = [
'์ค, ์ธ์ฌํ๋ค, ์์ฃผํ๋ค, ์น๋ค, ์ฌ๋, ๋ชฉํ',
'๊ทธ ์ค์ ์์ฃผํ ์ฌ๋์๊ฒ ๋ชฉํ์ ์น๋ฉฐ ์ธ์ฌ๋ฅผ ํ๋ค.',
'์ฌ์ฃผ๋ผ๋ ๊ฒ์ ์์ , ์์
, ํ์๊ธฐ, ๊ธ์ฐ๊ธฐ, ๋งํ๊ธฐ, ๊ณ์ฐํ๊ธฐ๋ค.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
๐ Documentation
Model Details
Model Description
Property | Details |
---|---|
Model Type | Sentence Transformer |
Base model | Alibaba-NLP/gte-multilingual-base |
Maximum Sequence Length | 8192 tokens |
Output Dimensionality | 768 tokens |
Similarity Function | Cosine Similarity |
Training Dataset | nlpai-lab/ko-triplet-v1.0 |
Language | ko |
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': 8192, 'do_lower_case': False}) with Transformer model: NewModel
(1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, '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': False, 'include_prompt': True})
(2): Normalize()
)
Evaluation
Metrics
Triplet
- Dataset:
dev
- Evaluated with
TripletEvaluator
Metric | Value |
---|---|
cosine_accuracy | 0.9855 |
dot_accuracy | 0.0145 |
manhattan_accuracy | 0.9855 |
euclidean_accuracy | 0.9855 |
max_accuracy | 0.9855 |
Training Details
Training Dataset
nlpai-lab/ko-triplet-v1.0
-
Dataset: nlpai-lab/ko-triplet-v1.0 at 9cc1d6a
-
Size: 10,000 training samples
-
Columns:
anchor
,positive
, andnegative
-
Approximate statistics based on the first 1000 samples:
anchor positive negative type string string string details - min: 9 tokens
- mean: 22.12 tokens
- max: 146 tokens
- min: 10 tokens
- mean: 92.69 tokens
- max: 1815 tokens
- min: 8 tokens
- mean: 99.24 tokens
- max: 880 tokens
-
Samples:
anchor positive negative ๊ธ์ฐ๊ธฐ, ๋ํ, ์ฐธ๊ฐ๋ฃ, ๋ฐ๋ค
๊ธ์ฐ๊ธฐ ๋ํ๋ ์ฐธ๊ฐ์ํํ ์ผ์ ๋ถ์ ์ฐธ๊ฐ๋ฃ๋ฅผ ๋ฐ์๋ค.
์ฌ์ฃผ๋ ์์ , ์์ , ํ์๊ธฐ, ๊ธ์ฐ๊ธฐ, ๋งํ๊ธฐ, ๊ณ์ฐํ๊ธฐ ๋ฑ๋ฑ ์ด๋ค.
"K๋ฆฌ๊ทธ 2002 ์์ฆ"์์ ๊ธฐ๋กํ ๊ด๊ฐ์ ์๋ณด๋ค ๋ ์ผ ๊ณ ํ๋ฆฌํ ๋ค์๋ฌ ๊ฒฝ๊ธฐ์ฅ์์ ๊ฐ์ต๋ "์ 4ํ ์ธ๊ณ ์ก์ ์ ์๊ถ ๋ํ"์ ๊ด๊ฐ ์๊ฐ ๋ง์๋?
1993๋ ์ธ๊ณ ์ก์ ์ ์๊ถ ๋ํ. ์ 4ํ ์ธ๊ณ ์ก์ ์ ์๊ถ ๋ํ๋ ๊ตญ์ ์ก์ ๊ฒฝ๊ธฐ ์ฐ๋งน ์ฃผ๊ด์ผ๋ก 1993๋ 8์ 13์ผ์์ 8์ 22์ผ๊น์ง ๋ ์ผ ์ํฌํธ๊ฐ๋ฅดํธ ๊ณ ํ๋ฆฌํ ๋ค์๋ฌ ๊ฒฝ๊ธฐ์ฅ์์ ์ด๋ฆฐ ๊ตญ์ ์ก์ ๋ํ์ด๋ค. ๋ ์ผ์์ ์ด๋ฆฐ ์ฒซ ๋ฒ์งธ ์ธ๊ณ ์ก์ ์ ์๊ถ ๋ํ์๋ค. 187๊ฐ๊ตญ ์ ์ 1630๋ช ์ด ์ฐธ๊ฐํ์ผ๋ฉฐ, ๋ํ ์ญ์ฌ์ ๊ฐ์ฅ ๋ง์ ์์ธ 58๋ง 5000๋ช ์ ๊ด์ค์ด ์ ์ฅํ๋ค.
์ก๊ฐํธ. ๊ฒฝ๋ ฅ. 1996-2000: ์ด๊ธฐ ๊ฒฝ๋ ฅ๊ณผ ์ฃผ๋ชฉ ๋ฐ๋ค. 1999๋ ์๋ ๊ฐ์ ๊ท ๊ฐ๋ ์ ์ํ ใ์ฌ๋ฆฌใ์ ์ด์ฅ๊ธธ ์ญํ ๋ก ์ถ์ฐํ๋ค. ์ด ์ํ๋ ๊ด๊ฐ์ 582๋ง ๋ช ์ ๊ธฐ๋กํ๋ฉฐ ๋น์ ๊ตญ๋ด ์ต๋ค ๊ด๊ฐ ์ํ ๊ธฐ๋ก์ ๊ฐฑ์ ํ๊ณ , ์ต์ด๋ก 500๋ง ๊ด๊ฐ์ ๋์ด์ฐ๋ค. 2000๋ ์๋ ์ฝ๋ฏธ๋ ์ํ ใ๋ฐ์น์ใ์์ ์ฒซ ์ฃผ์ฐ์ ๋งก์๋ค. ์ด ์ํ์์ ๊ทธ๋ ์ํ์์ด์ ๋ ์ฌ๋ฌ ๋ํธ ์ญํ ๋ก ์บ์คํ ๋์ด ์ํ๋ฅผ ์ํด ๋ ์ฌ๋ง ํ๋ จ์ ํ์๋ค. ์ดํ ์ก๊ฐํธ๋ ์ด ์ํ๊ฐ ๊ฐ์ฅ ์ธ์ ๊น๋ค๊ณ ๊ผฝ์ผ๋ฉฐ "๋ฌผ๋ฆฌ์ ์ผ๋ก ๊ฐ์ฅ ๊ทนํ์ ์ํฉ๊น์ง ๊ฐ ์ํ์ด์๋ค๋ ์๊ฐ์ด ๋ ๋ค. ๋ง์ฝ ์ง๊ธ ๋ ์ฌ๋ง์ ๋ค์ ํ๋ค๋ฉด ์ฃฝ์ ๊ฑฐ๋ค"๋ผ๊ณ ๋งํ๋ค. ๊ฐ์ ํด ๊ทธ๋ ๋ฐ์ฐฌ์ฑ ๊ฐ๋ ์ ์ํ ใ๊ณต๋๊ฒฝ๋น๊ตฌ์ญ JSAใ์์ ์ค๊ฒฝํ ์ค์ฌ ์ญํ ๋ก ์ถ์ฐํ๋ค. ์ด ์ํ๋ ์ด์ ๊น์ง ๋ฐ๊ณต์ด๋ฐ์ฌ๋ก๊ธฐ์์ ๋ฒ์ด๋์ง ๋ชปํ๋ ๋ถ๋จ ์์ฌ ํ๊ตญ์ํ์ ์๊ฐ์ ์ธ๊ฐ์ ๋ก ํ์ฅํ ์ํ์ผ๋ก 583๋ง ๋ช ์ ๊ด๊ฐ์๋ฅผ ๊ธฐ๋กํ๋ฉฐ ์ญ๋ ํฅํ 1์ ์ํ๋ก ๊ธฐ๋กํ๋ค. ์ก๊ฐํธ๋ 2019๋ ๋งค์ฒด์์ ์ธํฐ๋ทฐ์์ ๋ฐฐ์ฐ ์ธ์์ ์ ํ์ ์ด ๋๋ค๊ณ ์ธ๊ธํ๋ฉฐ โ๋ ์ํ๊ฐ ๊ฐ๋ดํ 2000๋ ์ ๋ฐฐ์ฐ ์ํ ์ด๋ฐ์ ๋ถ๊ธฐ์ ์ด ๋๋คโ๊ณ ๋งํ๋ค. ํ ์ํ ๊ด๊ณ์๋ ์ก๊ฐํธ์๊ฒ "ใ๋ฐ์น์ใ์์ ๋ณด์ฌ์ค๋ฐ์ด๋ ์ฐ๊ธฐ๋ฅผ ๋ณด๊ณ ๊ทธ ์ด์์ ์ฐ๊ธฐ๋ ๋์ฌ ์ ์์ ๊ฒ์ด๋ผ ์๊ฐํ๋ค. ์ด๋ฒ ์ํ๋ฅผ ๋ณด๊ณ ๋ด๊ฐ ๋น์ ์ฐ๊ธฐ์ ํ๊ณ๋ฅผ ๋๋ฌด ๋ฎ๊ฒ ์ก์์์ ์์๋ค"๊ณ ๋งํ๋ค. ใ๋งค์ผ๊ฒฝ์ ใ์ ํ๋ฆฌ๋ทฐ์์๋ "์ก๊ฐํธ๊ฐ ๊ทธ๋ ค๋ธ ๋ฐ๋ปํ๊ณ ๋๋ํ ์ค๊ฒฝํ ์ค์ฌ๋ ๊ตฐ๊ธฐ๊ฐ ํ์ฐธ ๋น ์ ธ ์๋ก ๋ ธ๋ฅ๊ฑฐ๋ฆฌ๋ ๊ฒ์ผ๋ก ๋น์น ์๋ ์๋ ๋ชจ์ต๋ค์ ๋๋ฌผ๋๋ ํ์ ์ ๋ก ์ก์์ฃผ๋ ๋ ๋ ํ ๋ฐ์นจ๋๋ค"๋ผ๊ณ ๋ฆฌ๋ทฐํ๋ค. ์ด๋ฌํ ํธํ ์์ ์ 1ํ ๋ถ์ฐ์ํํ๋ก ๊ฐํํ์, ์ 38ํ ๋์ข ์์ํ์ , ์ 3ํ ๋๋น์์์์ํ์ , ์์ ๋จ์ฐ์ฃผ์ฐ์์ ์์ํ๊ณ , ์ด ์ธ์๋ ๋ฐฑ์์์ ๋์์์ ์ธ๊ธฐ์์ ๋ฐ๋ ๋ฑ ๋ค์์ ์์์์์ ์ฐ๊ธฐ๋ ฅ๊ณผ ์คํ์ฑ์์ ์ธ์ ์ ๋ฐ์๋ค.
ํธ๋ฆฌ๊ฑฐ๋ ๋ฐ์ดํฐ๋ฅผ ์ ์ฅ์์ผ?
ํธ๋ฆฌ๊ฑฐ๋ ์ํ ๋ ์ํ ์ฅ๋น๋ฅผ ๊ตฌ๋์์ผ ์๋๋ฆฌ์ค๋ฅผ ๋ฐ์์ํค๋ฉฐ, ์๋๋ฆฌ์ค์ ๋ฐ๋ผ ์ํ ๋๋ถ์ ์ก์ ๊ธฐ์ ์์ ๊ธฐ๋ฅผ ์ ์ด ๋ฐ ์ธก์ ํ์ฌ ๋ฐ์ดํฐ๋ฅผ ์ ์ฅํ ๊ฒ ๋๋ค.
์ํ๊ด์์๋ ์ ๋๋ถ๋ถ์ ๊ด๊ฐ์ด ์ค๊ฐ๋๊ฒ ์๋ฆฌ๊ฐ ์ ๋ค๋ฆฌ๋ ๊ฑธ๊น? ๋์์ธ์ด Uglyํ์ง๋ง ์ CRT์ผ๋๋ ์๋ฆฌ๊ฐ ์ ๋ค๋ ธ์๊น? ์ด๋ฐ ์ง๋ฌธ์ ๋ํ ํด๋ต์ ์ํ๊ด ์คํฌ๋ฆฐ์์์ ์ฐพ์ ์ ์์๋ค.
-
Loss:
MultipleNegativesRankingLoss
with these parameters:{ "scale": 20.0, "similarity_fct": "cos_sim" }
Evaluation Dataset
nlpai-lab/ko-triplet-v1.0
- Dataset: nlpai-lab/ko-triplet-v1.0 at 9cc1d6a
- Size: 3,724 evaluation samples
- Columns:
anchor
,positive
, andnegative
- Approximate statistics based on the first 1000 samples:
anchor positive negative type string string string details - min: 6 tokens
- mean: 21.63 tokens
- max: 143 tokens
- min: 8 tokens
- mean: 88.89 tokens
- max: 2003 tokens
- min: 10 tokens
- mean: 102.66 tokens
- max: 3190 tokens
- Samples:
anchor positive negative
Jina Embeddings V3
Jina Embeddings V3 is a multilingual sentence embedding model supporting over 100 languages, specializing in sentence similarity and feature extraction tasks.
Text Embedding
Transformers Supports Multiple Languages

J
jinaai
3.7M
911
Ms Marco MiniLM L6 V2
Apache-2.0
A cross-encoder model trained on the MS Marco passage ranking task for query-passage relevance scoring in information retrieval
Text Embedding English
M
cross-encoder
2.5M
86
Opensearch Neural Sparse Encoding Doc V2 Distill
Apache-2.0
A sparse retrieval model based on distillation technology, optimized for OpenSearch, supporting inference-free document encoding with improved search relevance and efficiency over V1
Text Embedding
Transformers English

O
opensearch-project
1.8M
7
Sapbert From PubMedBERT Fulltext
Apache-2.0
A biomedical entity representation model based on PubMedBERT, optimized for semantic relation capture through self-aligned pre-training
Text Embedding English
S
cambridgeltl
1.7M
49
Gte Large
MIT
GTE-Large is a powerful sentence transformer model focused on sentence similarity and text embedding tasks, excelling in multiple benchmark tests.
Text Embedding English
G
thenlper
1.5M
278
Gte Base En V1.5
Apache-2.0
GTE-base-en-v1.5 is an English sentence transformer model focused on sentence similarity tasks, excelling in multiple text embedding benchmarks.
Text Embedding
Transformers Supports Multiple Languages

G
Alibaba-NLP
1.5M
63
Gte Multilingual Base
Apache-2.0
GTE Multilingual Base is a multilingual sentence embedding model supporting over 50 languages, suitable for tasks like sentence similarity calculation.
Text Embedding
Transformers Supports Multiple Languages

G
Alibaba-NLP
1.2M
246
Polybert
polyBERT is a chemical language model designed to achieve fully machine-driven ultrafast polymer informatics. It maps PSMILES strings into 600-dimensional dense fingerprints to numerically represent polymer chemical structures.
Text Embedding
Transformers

P
kuelumbus
1.0M
5
Bert Base Turkish Cased Mean Nli Stsb Tr
Apache-2.0
A sentence embedding model based on Turkish BERT, optimized for semantic similarity tasks
Text Embedding
Transformers Other

B
emrecan
1.0M
40
GIST Small Embedding V0
MIT
A text embedding model fine-tuned based on BAAI/bge-small-en-v1.5, trained with the MEDI dataset and MTEB classification task datasets, optimized for query encoding in retrieval tasks.
Text Embedding
Safetensors English
G
avsolatorio
945.68k
29
Featured Recommended AI Models
ยฉ 2025AIbase