gemma-molit-finetuned

Model Card for gemma-molit-finetuned

This model is a fine-tuned version of google/gemma-3-1b-it. It has been trained using TRL.

Quick start

from transformers import pipeline

question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="chohi/gemma-molit-finetuned", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])

Training procedure

This model was trained with SFT.

Framework versions

• TRL: 0.23.0
• Transformers: 4.57.1
• Pytorch: 2.8.0
• Datasets: 4.3.0
• Tokenizers: 0.22.1

Citations

Cite TRL as:

@misc{vonwerra2022trl,
	title        = {{TRL: Transformer Reinforcement Learning}},
	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallou{\'e}dec},
	year         = 2020,
	journal      = {GitHub repository},
	publisher    = {GitHub},
	howpublished = {\url{https://github.com/huggingface/trl}}
}

모델명: chohi/gemma-molit-finetuned

---

license: gemma
language:

• ko
• en
  base_model: google/gemma-3-1b-it tags:
• korean
• government
• molit
• rag
• domain-specific
• chatbot
• transportation
• public-sector pipeline_tag: text-generation

---

molit-gemma: 국토교통 도메인 특화 sLLM

molit-gemma는 Google Gemma-3-1B를 대한민국 국토교통부(MOLIT) 도메인 데이터로
파인튜닝한 한국어 정부 정책 질의응답 특화 sLLM입니다.

Model Description

• Base model: google/gemma-3-1b-it
• Domain: 국토교통(주택, 도로, 교통, 부동산 정책)
• Use case: 정부 민원 질의응답, 정책 문서 기반 RAG 챗봇
• Deployment: 온프레미스(공공기관 보안 환경)

Performance

Architecture

• Retrieval: OpenSearch 기반 정책 문서 검색
• Generation: molit-gemma (fine-tuned Gemma-3-1B)
• RAG로 환각(hallucination) 위험 완화

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
                                                                                                                                                         
model = AutoModelForCausalLM.from_pretrained("chohi/gemma-molit-finetuned")
tokenizer = AutoTokenizer.from_pretrained("chohi/gemma-molit-finetuned")                                                                                
                                                                                                                                                         
prompt = "신혼부부 특별공급 자격 조건은?"
inputs = tokenizer(prompt, return_tensors="pt")                                                                                                          
outputs = model.generate(**inputs, max_new_tokens=512)
print(tokenizer.decode(outputs[0]))                                                                                                                      

Citation

정현일. (2026). 거대언어모델 기반 국토교통 분야 데이터 질의응답 챗봇 시스템 설계 및 구현 [석사학위논문, 충북대학교].
RISS: https://www.riss.kr/link?id=T17378943

Limitations

• 국토교통 외 도메인 성능 저하
• 1B 파라미터 한계로 복잡한 추론 제한
• 학습 데이터 시점 이후 정책 변경 미반영 (RAG로 보완 권장)

License

Gemma Terms of Use 준수