【Langchain大语言模型开发教程】记忆

🔗 LangChain for LLM Application Development - DeepLearning.AI

学习目标

1、Langchain的历史记忆 ConversationBufferMemory

2、基于窗口限制的临时记忆 ConversationBufferWindowMemory

3、基于Token数量的临时记忆 ConversationTokenBufferMemory

4、基于历史内容摘要的临时记忆 ConversationSummaryMemory

Langchain的历史记忆（ConversationBufferMemory）

import os
import warnings
from dotenv import load_dotenv, find_dotenv
from langchain.chat_models import ChatOpenAI
from langchain.chains import ConversationChain
from langchain.memory import ConversationBufferMemory

_ = load_dotenv(find_dotenv())
warnings.filterwarnings('ignore')

我们依然使用智谱的LLM，实例化一下Langchain的记忆模块，并构建一个带有记忆的对话模型

llm = ChatOpenAI(api_key=os.environ.get('ZHIPUAI_API_KEY'),
                         base_url=os.environ.get('ZHIPUAI_API_URL'),
                         model="glm-4",
                         temperature=0.98)

memory = ConversationBufferMemory()
conversation = ConversationChain(
    llm=llm,
    memory = memory,
    verbose=True
)

进行对话

conversation.predict(input="Hi, my name is Andrew")
conversation.predict(input="What is 1+1?")
conversation.predict(input="What is my name?")

模型确实可以记住我们的名字,打印一下记忆内容

#两种方式
print(memory.buffer)

memory.load_memory_variables({})

此外，Langchain还提供了一个函数来添加对话内容

memory.save_context({"input": "Hi"},
                    {"output": "What's up"})

基于窗口限制的临时记忆（ConversationBufferWindowMemory）

from langchain.memory import ConversationBufferWindowMemory

memory = ConversationBufferWindowMemory(k=1) #k表示我们保留最近几轮对话的数量

我们先来添加两轮对话

memory.save_context({"input": "Hi"},
                    {"output": "What's up"})
memory.save_context({"input": "Not much, just hanging"},
                    {"output": "Cool"})

通过对话历史可以发现，记忆中只保存了一轮的信息

memory.load_memory_variables({})

{'history': 'Human: Not much, just hanging\nAI: Cool'}

我们使用这种记忆方式来构建一个对话模型，发现他确实遗忘了第一轮的信息

llm = ChatOpenAI(api_key=os.environ.get('ZHIPUAI_API_KEY'),
                         base_url=os.environ.get('ZHIPUAI_API_URL'),
                         model="glm-4",
                         temperature=0.98)
memory = ConversationBufferWindowMemory(k=1)
conversation = ConversationChain(
    llm=llm,
    memory = memory,
    verbose=False
)

基于Token数量的临时记忆 ConversationTokenBufferMemory

由于langchain中计算token数量的函数并不支持GLM4，所有使用这个函数会报错,根据源代码目前是支持gpt-3.5-turbo-0301、gpt-3.5-turbo、gpt-4，不知道以后会不会加入国产的这些模型。

memory = ConversationTokenBufferMemory(llm=llm, max_token_limit=500)

memory.save_context({"input": "AI is what?!"},
                    {"output": "Amazing!"})
memory.save_context({"input": "Backpropagation is what?"},
                    {"output": "Beautiful!"})
memory.save_context({"input": "Chatbots are what?"}, 
                    {"output": "Charming!"})

基于历史内容摘要的临时记忆 ConversationSummaryMemory

同理哈，这个函数的作用就是，我们会将历史的对话信息进行总结然后存在我们的记忆单元中，由于这里同样涉及到token的计算，所以这里也是无法正常运行的了。

from langchain.memory import ConversationSummaryBufferMemory
# create a long string
schedule = "There is a meeting at 8am with your product team. \
You will need your powerpoint presentation prepared. \
9am-12pm have time to work on your LangChain \
project which will go quickly because Langchain is such a powerful tool. \
At Noon, lunch at the italian resturant with a customer who is driving \
from over an hour away to meet you to understand the latest in AI. \
Be sure to bring your laptop to show the latest LLM demo."

memory = ConversationSummaryBufferMemory(llm=llm, max_token_limit=100)
memory.save_context({"input": "Hello"}, {"output": "What's up"})
memory.save_context({"input": "Not much, just hanging"},
                    {"output": "Cool"})
memory.save_context({"input": "What is on the schedule today?"},
                    {"output": f"{schedule}"})

构建一个对话模型 （verbose设置为true可以查看到我们历史的一些信息）

conversation = ConversationChain(
    llm=llm,
    memory = memory,
    verbose=True
)

尝试进行提问

conversation.predict(input="What would be a good demo to show?")

总结（吴恩达老师视频中的内容）