mac

安裝

• pip install mlx-lmCopy

使用 mac cli

• python -m mlx_lm.generate --model mlx-community/Phi-3-mini-4k-instruct-4bit --prompt "how are you"Copy

import bs4

from langchain import hub

from langchain_community.document_loaders import WebBaseLoader

from langchain_chroma import Chroma

from langchain_core.output_parsers import StrOutputParser

from langchain_core.runnables import RunnablePassthrough

from langchain_openai import OpenAIEmbeddings

from langchain_text_splitters import RecursiveCharacterTextSplitter
Copy

import getpass

import os

  

os.environ["OPENAI_API_KEY"] = "sk-proj-"#getpass.getpass()
from langchain_community.document_loaders import PyPDFLoader

  

loader = PyPDFLoader("files/替代役管理作業規定.pdf")

pages = loader.load_and_split()

Copy

from langchain_community.llms.mlx_pipeline import MLXPipeline

  

llm = MLXPipeline.from_model_id(

"mlx-community/quantized-gemma-2b-it",

# pipeline_kwargs={"max_tokens": 512, "temp": 0.1},

pipeline_kwargs={"temp": 0.1},

)
Copy

from huggingface_hub import login

login(token="")
Copy

import os

  
  

docs = loader.load()

  

text_splitter = RecursiveCharacterTextSplitter(chunk_size=1000, chunk_overlap=200)

splits = text_splitter.split_documents(docs)

vectorstore = Chroma.from_documents(documents=splits, embedding=OpenAIEmbeddings())

  

# Retrieve and generate using the relevant snippets of the blog.

retriever = vectorstore.as_retriever()

prompt = hub.pull("rlm/rag-prompt")

  

def format_docs(docs):

return "\n\n".join(doc.page_content for doc in docs)

  
  

rag_chain = (

{"context": retriever | format_docs, "question": RunnablePassthrough()}

| prompt

| llm

| StrOutputParser()

)
Copy

question = "替代役管理作業規定是依據甚麼法律?"

rag_chain.invoke(f"{question} ")
# ' 根據上述法律，替代役管理作業規定主要依據兵役法第 26條。'
Copy

Ollama

• 使用 Ollama 執行 TAIDE 的 TAIDE-LX-7B-Chat-4bit 大語言模型
  • ollama create taide-lx-7b-chat-4bit:latest -f ModelfileCopy
• 介紹好用工具：Ollama 快速在本地啟動並執行大型語言模型 透過Ollama在本機電腦執行大型語言模型（LLM）：Windows與VS Code篇

Ref