README.md

metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - telepix
pipeline_tag: sentence-similarity
library_name: sentence-transformers

PIXIE-Rune-v1.0

PIXIE-Rune-v1.0 is an encoder-based embedding model trained on Korean and English triplets, developed by TelePIX Co., Ltd. PIXIE stands for TelePIX Intelligent Embedding, representing TelePIX’s high-performance embedding technology. The model is multilingual, specifically optimized for both Korean and English. It demonstrates strong performance on retrieval tasks in both languages, achieving robust results across a wide range of Korean- and English-language benchmarks. This makes it well-suited for real-world applications that require high-quality semantic search in Korean, English, or both.

Model Description

• Model Type: Sentence Transformer
• Maximum Sequence Length: 8192 tokens
• Output Dimensionality: 1024 dimensions
• Similarity Function: Cosine Similarity
• Language: Bilingual — optimized for high performance in Korean and English
• License: apache-2.0

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 8192, 'do_lower_case': False}) with Transformer model: XLMRobertaModel 
  (1): Pooling({'word_embedding_dimension': 1024, '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()
)

Quality Benchmarks

PIXIE-Rune-v1.0 is a multilingual embedding model specialized for Korean and English retrieval tasks. It delivers consistently strong performance across a diverse set of domain-specific and open-domain benchmarks in both languages, demonstrating its effectiveness in real-world semantic search applications. The table below presents the retrieval performance of several embedding models evaluated on a variety of Korean and English benchmarks. We report Normalized Discounted Cumulative Gain (NDCG) scores, which measure how well a ranked list of documents aligns with ground truth relevance. Higher values indicate better retrieval quality.

• Avg. NDCG: Average of NDCG@1, @3, @5, and @10 across all benchmark datasets.
• NDCG@k: Relevance quality of the top-k retrieved results.

Korean Retrieval Benchmarks

Our model, telepix/PIXIE-Rune-v1.0, achieves state-of-the-art performance across most metrics and benchmarks, demonstrating strong generalization across domains such as multi-hop QA, long-document retrieval, public health, and e-commerce.

Model Name	# params	Avg. NDCG	NDCG@1	NDCG@3	NDCG@5	NDCG@10
telepix/PIXIE-Rune-v1.0	568M	0.6905	0.6461	0.6859	0.7063	0.7238
nlpai-lab/KURE-v1	568M	0.6751	0.6277	0.6725	0.6907	0.7095
dragonekue/BGE-m3-ko	568M	0.6658	0.6225	0.6627	0.6795	0.6985
Snowflake/snowflake-arctic-embed-l-v2.0	568M	0.6592	0.6118	0.6542	0.6759	0.6949
BAAI/bge-m3	568M	0.6573	0.6099	0.6533	0.6732	0.6930
Qwen/Qwen3-Embedding-0.6B	595M	0.6321	0.5894	0.6274	0.6455	0.6662
jinaai/jina-embeddings-v3	572M	0.6293	0.5800	0.6254	0.6456	0.6665
Alibaba-NLP/gte-multilingual-base	305M	0.6111	0.5542	0.6089	0.6302	0.6511
openai/text-embedding-3-large	N/A	0.6015	0.5466	0.5999	0.6187	0.6409

Descriptions of the benchmark datasets used for evaluation are as follows:

• Ko-StrategyQA
  A Korean multi-hop open-domain question answering dataset designed for complex reasoning over multiple documents.
• AutoRAGRetrieval
  A domain-diverse retrieval dataset covering finance, government, healthcare, legal, and e-commerce sectors.
• MIRACLRetrieval
  A document retrieval benchmark built on Korean Wikipedia articles.
• PublicHealthQA
  A retrieval dataset focused on medical and public health topics.
• BelebeleRetrieval
  A dataset for retrieving relevant content from web and news articles in Korean.
• MultiLongDocRetrieval
  A long-document retrieval benchmark based on Korean Wikipedia and mC4 corpus.
• XPQARetrieval
  A real-world dataset constructed from user queries and relevant product documents in a Korean e-commerce platform.

English Retrieval Benchmarks

Our model, telepix/PIXIE-Rune-v1.0, achieves strong performance on a wide range of tasks, including fact verification, multi-hop question answering, financial QA, and scientific document retrieval, demonstrating competitive generalization across diverse domains.

Model Name	# params	Avg. NDCG	NDCG@1	NDCG@3	NDCG@5	NDCG@10
telepix/PIXIE-Rune-v1.0	568M	0.5781	0.5691	0.5663	0.5791	0.5979
Snowflake/snowflake-arctic-embed-l-v2.0	568M	0.5812	0.5725	0.5705	0.5811	0.6006
Qwen/Qwen3-Embedding-0.6B	595M	0.5558	0.5321	0.5451	0.5620	0.5839
Alibaba-NLP/gte-multilingual-base	305M	0.5541	0.5446	0.5426	0.5574	0.5746
BAAI/bge-m3	568M	0.5318	0.5078	0.5231	0.5389	0.5573
dragonekue/BGE-m3-ko	568M	0.5307	0.5125	0.5174	0.5362	0.5566
nlpai-lab/KURE-v1	568M	0.5272	0.5017	0.5171	0.5353	0.5548

Descriptions of the benchmark datasets used for evaluation are as follows:

• ArguAna
  A dataset for argument retrieval based on claim-counterclaim pairs from online debate forums.
• FEVER
  A fact verification dataset using Wikipedia for evidence-based claim validation.
• FiQA-2018
  A retrieval benchmark tailored to the finance domain with real-world questions and answers.
• HotpotQA
  A multi-hop open-domain QA dataset requiring reasoning across multiple documents.
• MSMARCO
  A large-scale benchmark using real Bing search queries and corresponding web documents.
• NQ
  A Google QA dataset where user questions are answered using Wikipedia articles.
• SCIDOCS
  A citation-based document retrieval dataset focused on scientific papers.

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
  
# Load the model
model_name = 'PIXIE-Rune-v1.0'
model = SentenceTransformer(model_name)

# Define the queries and documents
queries = [
    "텔레픽스는 어떤 산업 분야에서 위성 데이터를 활용하나요?",
    "국방 분야에 어떤 위성 서비스가 제공되나요?",
    "텔레픽스의 기술 수준은 어느 정도인가요?",
]
documents = [
    "텔레픽스는 국방, 농업, 자원, 해양 등 다양한 분야에서 위성 데이터를 분석하여 서비스를 제공합니다.",
    "정찰 및 감시 목적의 위성 영상을 통해 국방 관련 정밀 분석 서비스를 제공합니다.",
    "TelePIX의 광학 탑재체 및 AI 분석 기술은 Global standard를 상회하는 수준으로 평가받고 있습니다.",
    "텔레픽스는 우주에서 수집한 정보를 분석하여 '우주 경제(Space Economy)'라는 새로운 가치를 창출하고 있습니다.",
    "텔레픽스는 위성 영상 획득부터 분석, 서비스 제공까지 전 주기를 아우르는 솔루션을 제공합니다.",
]

# Compute embeddings: use `prompt_name="query"` to encode queries!
query_embeddings = model.encode(queries, prompt_name="query")
document_embeddings = model.encode(documents)

# Compute cosine similarity scores
scores = model.similarity(query_embeddings, document_embeddings)

# Output the results
for query, query_scores in zip(queries, scores):
    doc_score_pairs = list(zip(documents, query_scores))
    doc_score_pairs = sorted(doc_score_pairs, key=lambda x: x[1], reverse=True)
    print("Query:", query)
    for document, score in doc_score_pairs:
        print(score, document)

Framework Versions

• Python: 3.10.16
• Sentence Transformers: 4.0.1
• Transformers: 4.51.3
• PyTorch: 2.6.0+cu124
• Accelerate: 1.5.2
• Datasets: 2.21.0
• Tokenizers: 0.21.1

Contact

If you have any suggestions or questions about this Model, please reach out to the authors at bmkim@telepix.net.