Skip to main content
이 노트북은 MLX LLM을 채팅 모델로 사용하는 방법을 보여줍니다. 특히 다음을 수행합니다:
  1. MLXPipeline을 활용합니다.
  2. ChatMLX 클래스를 활용하여 이러한 LLM들이 LangChain의 Chat Messages 추상화와 인터페이스할 수 있도록 합니다.
  3. 오픈소스 LLM을 사용하여 ChatAgent 파이프라인을 구동하는 방법을 시연합니다.
pip install -qU  mlx-lm transformers huggingface_hub

​
1. LLM 인스턴스화

선택할 수 있는 세 가지 LLM 옵션이 있습니다. 
from langchain_community.llms.mlx_pipeline import MLXPipeline

llm = MLXPipeline.from_model_id(
    "mlx-community/quantized-gemma-2b-it",
    pipeline_kwargs={"max_tokens": 10, "temp": 0.1},
)

​
2. 채팅 템플릿을 적용하기 위한 ChatMLX 인스턴스화

채팅 모델과 전달할 메시지들을 인스턴스화합니다. 
from langchain_community.chat_models.mlx import ChatMLX
from langchain.messages import HumanMessage

messages = [
    HumanMessage(
        content="What happens when an unstoppable force meets an immovable object?"
    ),
]

chat_model = ChatMLX(llm=llm)
LLM 호출을 위해 채팅 메시지가 어떻게 포맷되는지 확인합니다. 
chat_model._to_chat_prompt(messages)
모델을 호출합니다. 
res = chat_model.invoke(messages)
print(res.content)

​
3. 에이전트로 테스트하기

여기서는 gemma-2b-it을 제로샷 ReAct 에이전트로 테스트합니다. 아래 예제는 여기에서 가져온 것입니다.
참고: 이 섹션을 실행하려면 SerpAPI Token을 환경 변수 SERPAPI_API_KEY로 저장해야 합니다. 
from langchain_classic import hub
from langchain.agents import AgentExecutor, load_tools
from langchain.agents.format_scratchpad import format_log_to_str
from langchain.agents.output_parsers import (
    ReActJsonSingleInputOutputParser,
)
from langchain.tools.render import render_text_description
from langchain_community.utilities import SerpAPIWrapper
react-json 스타일 프롬프트와 검색 엔진 및 계산기에 대한 액세스로 에이전트를 구성합니다. 
# setup tools
tools = load_tools(["serpapi", "llm-math"], llm=llm)

# setup ReAct style prompt
# Based on 'hwchase17/react' prompt modification, cause mlx does not support the `System` role
human_prompt = """
Answer the following questions as best you can. You have access to the following tools:

{tools}

The way you use the tools is by specifying a json blob.
Specifically, this json should have a `action` key (with the name of the tool to use) and a `action_input` key (with the input to the tool going here).

The only values that should be in the "action" field are: {tool_names}

The $JSON_BLOB should only contain a SINGLE action, do NOT return a list of multiple actions. Here is an example of a valid $JSON_BLOB:

\`\`\`
{{
  "action": $TOOL_NAME,
  "action_input": $INPUT
}}
\`\`\`

ALWAYS use the following format:

Question: the input question you must answer
Thought: you should always think about what to do
Action:
\`\`\`
$JSON_BLOB
\`\`\`
Observation: the result of the action
... (this Thought/Action/Observation can repeat N times)
Thought: I now know the final answer
Final Answer: the final answer to the original input question

Begin! Reminder to always use the exact characters `Final Answer` when responding.

{input}

{agent_scratchpad}

"""

prompt = human_prompt.partial(
    tools=render_text_description(tools),
    tool_names=", ".join([t.name for t in tools]),
)

# define the agent
chat_model_with_stop = chat_model.bind(stop=["\nObservation"])
agent = (
    {
        "input": lambda x: x["input"],
        "agent_scratchpad": lambda x: format_log_to_str(x["intermediate_steps"]),
    }
    | prompt
    | chat_model_with_stop
    | ReActJsonSingleInputOutputParser()
)

# instantiate AgentExecutor
agent_executor = AgentExecutor(agent=agent, tools=tools, verbose=True)

agent_executor.invoke(
    {
        "input": "Who is Leo DiCaprio's girlfriend? What is her current age raised to the 0.43 power?"
    }
)
Edit the source of this page on GitHub.
Connect these docs programmatically to Claude, VSCode, and more via MCP for real-time answers.
⌘I