成品网站nike源码1688腾讯企业qq官网

Large Language Models (LLMs) 在语义知识方面表现不错，但也有一些不足，如：不能正确计算数学公式、无法获取最新知识新闻

通过 Agents 可以赋予 LLMs 更多能力，让LLM能够计算、上网查询

agent 简单使用

from langchain import OpenAI
# 语言模型
llm = OpenAI(
openai_api_key="OPENAI_API_KEY",
temperature=0,
model_name="text-davinci-003"
)from langchain.chains import LLMMathChain
from langchain.agents import Tool
# 能计算数学公式的一个chain
llm_math = LLMMathChain(llm=llm)# initialize the math tool
math_tool = Tool(
name='Calculator',
func=llm_math.run,
description='Useful for when you need to answer questions about math.' # 描述工具能做什么
)
# when giving tools to LLM, we must pass as list of tools
tools = [math_tool]# 如果 langchain.agents 中有相关工具，则可以直接使用
#from langchain.agents import load_tools
#tools = load_tools(
#['llm-math'],
#llm=llm
)# 初始化 agent
from langchain.agents import initialize_agent
zero_shot_agent = initialize_agent(agent="zero-shot-react-description", # 无记忆的agenttools=tools, # tools 中只有math_tool，所以只能做计算llm=llm,verbose=True, # 显示执行过程max_iterations=3)
zero_shot_agent("what is (4.5*2.1)^2.2?")

上面的 tools 中只有math_tool，所以 zero_shot_agent 只能做计算，不能回答其它常识问题，可以在 tools 中添加更多工具，使得 zero_shot_agent 拥有更多能力。

# 可以在 tools 中新增聊天工具
from langchain.prompts import PromptTemplate
from langchain.chains import LLMChain
prompt = PromptTemplate(
input_variables=["query"],
template="{query}"
)
llm_chain = LLMChain(llm=llm, prompt=prompt)# initialize the LLM tool
llm_tool = Tool(
name='Language Model',
func=llm_chain.run,
description='use this tool for general purpose queries and logic'
)
tools.append(llm_tool)
# reinitialize the agent
zero_shot_agent = initialize_agent(
agent="zero-shot-react-description",
tools=tools,
llm=llm,
verbose=True,
max_iterations=3
)

agent 类型

zero-shot-react-description 无缓存的方式，聊天是单次的，无上下文缓存

zero_shot_agent = initialize_agent(
agent="zero-shot-react-description",
tools=tools,
llm=llm,
verbose=True,
max_iterations=3,
)

conversational-react-description 带缓存

from langchain.memory import ConversationBufferMemorymemory = ConversationBufferMemory(memory_key="chat_history")conversational_agent = initialize_agent(
agent='conversational-react-description',
tools=tools,
llm=llm,
verbose=True,
max_iterations=3,
memory=memory,
)

react-docstore 可以检索知识库，无缓存

from langchain import Wikipedia
from langchain.agents.react.base import DocstoreExplorerdocstore=DocstoreExplorer(Wikipedia())
tools = [Tool(name="Search", # 信息检索func=docstore.search, description='search wikipedia'),Tool(name="Lookup", # 匹配相近结果func=docstore.lookup, description='lookup a term in wikipedia')
]docstore_agent = initialize_agent(tools,llm,agent="react-docstore",verbose=True,max_iterations=3)

self-ask-with-search 将LLM与搜索引擎结合起来

from langchain import SerpAPIWrapper# initialize the search chain
search = SerpAPIWrapper(serpapi_api_key='serp_api_key')# create a search tool
tools = [Tool(name="Intermediate Answer",func=search.run,description='google search')]# initialize the search enabled agent
self_ask_with_search = initialize_agent(tools,llm,agent="self-ask-with-search",verbose=True)

参考：
Superpower LLMs with Conversational Agents