로컬 LLM 완전 가이드: Ollama로 GPT급 AI를 내 PC에서 실행하기 (2025)

# macOS
brew install ollama

# Linux
curl -fsSL https://ollama.ai/install.sh | sh

# 모델 다운로드 및 실행
ollama run llama3.2        # Meta Llama 3.2 (3B, ~2GB)
ollama run llama3.1:8b     # Llama 3.1 8B (~5GB)
ollama run qwen2.5:7b      # Alibaba Qwen 2.5 7B (~5GB)
ollama run gemma2:9b       # Google Gemma 2 9B (~6GB)
ollama run mistral         # Mistral 7B (~4GB)

from openai import OpenAI

# 기존: client = OpenAI(api_key="sk-...")
client = OpenAI(base_url="http://localhost:11434/v1", api_key="ollama")

response = client.chat.completions.create(
    model="llama3.1:8b",
    messages=[{"role": "user", "content": "안녕하세요!"}]
)
print(response.choices[0].message.content)

ollama run qwen2.5:14b  # 14B, ~9GB, 한국어 최강

# 빌드
git clone https://github.com/ggerganov/llama.cpp
cd llama.cpp && make -j4

# Apple Silicon Metal 가속
LLAMA_METAL=1 make -j4

# CUDA 가속 (NVIDIA)
LLAMA_CUDA=1 make -j4

# 실행
./main -m ./models/llama-3.1-8b-q4_k_m.gguf        -n 512 --temp 0.7 -p "한국의 AI 트렌드를 분석해줘"

import ollama

def review_code(code: str) -> str:
    response = ollama.chat(
        model="qwen2.5:14b",
        messages=[{
            "role": "user",
            "content": f"다음 코드를 리뷰해주세요. 보안 취약점, 성능 이슈, 개선 사항을 지적해주세요:

{code}"
        }]
    )
    return response["message"]["content"]

# LlamaIndex + Ollama 조합
from llama_index.core import VectorStoreIndex, SimpleDirectoryReader
from llama_index.llms.ollama import Ollama
from llama_index.embeddings.ollama import OllamaEmbedding

llm = Ollama(model="llama3.1:8b", request_timeout=120.0)
embed_model = OllamaEmbedding(model_name="nomic-embed-text")

documents = SimpleDirectoryReader("./internal_docs").load_data()
index = VectorStoreIndex.from_documents(documents, embed_model=embed_model)
query_engine = index.as_query_engine(llm=llm)

response = query_engine.query("우리 회사 휴가 정책이 어떻게 돼?")