LM Studio

LM Studio

LM Studio로 로컬 LLM을 발견·다운로드·실행할 수 있어요. 모든 LM Studio 모델을 지원해요. litellm 요청 시 model=lm_studio/<any-model-on-lmstudio> 접두사로 설정하기만 하면 돼요.

출처: 문서

본문

개요 (Overview)

API 키

# env variable
os.environ['LM_STUDIO_API_BASE']
os.environ['LM_STUDIO_API_KEY']  # optional, default is empty

샘플 사용법 (Sample Usage)

from litellm import completion
import os

os.environ['LM_STUDIO_API_BASE'] = ""

response = completion(
    model="lm_studio/llama-3-8b-instruct",
    messages=[
        {
            "role": "user",
            "content": "What's the weather like in Boston today in Fahrenheit?",
        }
    ]
)
print(response)

샘플 사용법 - 스트리밍 (Streaming)

from litellm import completion
import os

os.environ['LM_STUDIO_API_KEY'] = ""

response = completion(
    model="lm_studio/llama-3-8b-instruct",
    messages=[
        {
            "role": "user",
            "content": "What's the weather like in Boston today in Fahrenheit?",
        }
    ],
    stream=True,
)

for chunk in response:
    print(chunk)

LiteLLM Proxy Server 사용법

config.yaml 수정:

model_list:
  - model_name: my-model
    litellm_params:
      model: lm_studio/<your-model-name>  # add lm_studio/ prefix to route as LM Studio provider
      api_key: api-key  # api key to send your model

Proxy 시작:

$ litellm --config /path/to/config.yaml

LiteLLM Proxy Server로 요청:

OpenAI Python v1.0.0+:

import openai

client = openai.OpenAI(
    api_key="sk-<your-litellm-api-key>",  # pass litellm proxy key, if you're using virtual keys
    base_url="http://0.0.0.0:4000"  # litellm-proxy-base url
)

response = client.chat.completions.create(
    model="my-model",
    messages = [
        {
            "role": "user",
            "content": "what llm are you"
        }
    ],
)
print(response)

curl:

curl --location 'http://0.0.0.0:4000/chat/completions' \
  --header "Authorization: Bearer ***" \
  --header 'Content-Type: application/json' \
  --data '{
    "model": "my-model",
    "messages": [
      {
        "role": "user",
        "content": "what llm are you"
      }
    ],
  }'

지원 파라미터 (Supported Parameters)

지원 파라미터 목록은 Supported Parameters 문서를 참고하세요.

Embedding

from litellm import embedding
import os

os.environ['LM_STUDIO_API_BASE'] = "http://localhost:8000"

response = embedding(
    model="lm_studio/jina-embeddings-v3",
    input=["Hello world"],
)
print(response)

구조화 출력 (Structured Output)

Pydantic 모델을 전달해 구조화된 출력을 받을 수 있어요.

from pydantic import BaseModel
from litellm import completion

class Book(BaseModel):
    title: str
    author: str
    year: int

response = completion(
    model="lm_studio/llama-3-8b-instruct",
    messages=[{"role": "user", "content": "Tell me about The Hobbit"}],
    response_format=Book,
)
print(response.choices[0].message.content)

더 알아보기 (Learn more)

• LiteLLM 개요
• LiteLLM Python SDK 사용법
• LiteLLM Proxy Server 설정
• 지원되는 공급업체 전체 목록