LangChain Memory
TIP
Memory 让 LLM 记住对话历史，实现多轮对话的上下文感知。
ConversationBufferMemory
python
from langchain.memory import ConversationBufferMemory
memory = ConversationBufferMemory()
memory.chat_memory.add_user_message("你好!")
memory.chat_memory.add_ai_message("你好！有什么可以帮你的吗？")
memory.load_memory_variables({})ConversationSummaryMemory
python
from langchain.memory import ConversationSummaryMemory
memory = ConversationSummaryMemory(
llm=ChatOpenAI(model="gpt-4"),
max_token_limit=200
)
memory.chat_memory.add_user_message("我叫张三")
memory.chat_memory.add_ai_message("你好张三！")在 Chain 中使用
python
from langchain.chains import ConversationChain
conversation = ConversationChain(
llm=ChatOpenAI(temperature=0),
memory=ConversationBufferMemory()
)
conversation.predict(input="你好！我是张三")
conversation.predict(input="你还记得我的名字吗？")
# AI: 当然记得！