LLM RAG with Agent
发布时间
阅读量：
阅读量
题意： 基于代理的LLM检索增强生成
问题背景：
I have been working on the application code located at this link: https://python.langchain.com/docs/use_cases/question_answering/conversational_retrieval_agents. By engaging with the application code, I have been practicing and analyzing its functionality to enhance my proficiency in Python and AI applications.
我正在尝试链接中的应用程序代码。
I am using the following Llang Chain version
我正在使用以下Llang Chain版本
langchain 0.0.327 langchain-community 0.0.2 langchain-core 0.1.0
Getting the following error: 得到以下的错误
Entering new AgentExecutor chain...
Traceback (most recent call last):
File "RAGWithAgent.py", line 54, in <module>
result = agent_executor({"input": "hi, im bob"})
File "\lib\site-packages\langchain\chains\base.py", line 310, in __call__
raise e
File "\lib\site-packages\langchain\chains\base.py", line 304, in __call__
self._call(inputs, run_manager=run_manager)
File "\lib\site-packages\langchain\agents\agent.py", line 1146, in _call
next_step_output = self._take_next_step(
File "\lib\site-packages\langchain\agents\agent.py", line 933, in _take_next_step
output = self.agent.plan(
File "\lib\site-packages\langchain\agents\openai_functions_agent\base.py", line 104, in plan
predicted_message = self.llm.predict_messages(
File "\lib\site-packages\langchain\chat_models\base.py", line 650, in predict_messages
return self(messages, stop=_stop, **kwargs)
File "\lib\site-packages\langchain\chat_models\base.py", line 600, in __call__
generation = self.generate(
File "\lib\site-packages\langchain\chat_models\base.py", line 349, in generate
raise e
File "\lib\site-packages\langchain\chat_models\base.py", line 339, in generate
self._generate_with_cache(
File "\lib\site-packages\langchain\chat_models\base.py", line 492, in _generate_with_cache
return self._generate(
File "\lib\site-packages\langchain\chat_models\openai.py", line 357, in _generate
return _generate_from_stream(stream_iter)
File "\lib\site-packages\langchain\chat_models\base.py", line 57, in _generate_from_stream
for chunk in stream:
File "\lib\site-packages\langchain\chat_models\openai.py", line 326, in _stream
for chunk in self.completion_with_retry(
File "\lib\site-packages\langchain\chat_models\openai.py", line 299, in completion_with_retry
return _completion_with_retry(**kwargs)
File "\lib\site-packages\tenacity\__init__.py", line 289, in wrapped_f
return self(f, *args, **kw)
File "\lib\site-packages\tenacity\__init__.py", line 379, in __call__
do = self.iter(retry_state=retry_state)
File "\lib\site-packages\tenacity\__init__.py", line 314, in iter
return fut.result()
File "D:\Program Files\Python38\lib\concurrent\futures\_base.py", line 432, in result
return self.__get_result()
File "D:\Program Files\Python38\lib\concurrent\futures\_base.py", line 388, in __get_result
raise self._exception
File "\lib\site-packages\tenacity\__init__.py", line 382, in __call__
result = fn(*args, **kwargs)
File "\lib\site-packages\langchain\chat_models\openai.py", line 297, in _completion_with_retry
return self.client.create(**kwargs)
File "\lib\site-packages\openai\api_resources\chat_completion.py", line 25, in create
return super().create(*args, **kwargs)
File "\lib\site-packages\openai\api_resources\abstract\engine_api_resource.py", line 155, in create
response, _, api_key = requestor.request(
File "\lib\site-packages\openai\api_requestor.py", line 299, in request
resp, got_stream = self._interpret_response(result, stream)
File "\lib\site-packages\openai\api_requestor.py", line 710, in _interpret_response
self._interpret_response_line(
File "\lib\site-packages\openai\api_requestor.py", line 775, in _interpret_response_line
raise self.handle_error_response(
openai.error.InvalidRequestError: Unrecognized request argument supplied: functions
Process finished with exit code 1
代码解读Instead of using openAI, I employed Azure LLM. Since FAISS wasn't performing well for me, I switched to using Chroma Vector Store instead.
我选择了Azure的LLM而非OpenAI。由于FAISS不起作用的原因,因此我采用了Chroma Vector Store.
Following is my code: 下面是我的代码
from langchain.text_splitter import CharacterTextSplitter
from langchain.document_loaders import TextLoader
from langchain.agents.agent_toolkits import create_retriever_tool
from langchain.agents.agent_toolkits import create_conversational_retrieval_agent
from langchain.chat_models import AzureChatOpenAI
from langchain.vectorstores import Chroma
from langchain.embeddings.sentence_transformer import SentenceTransformerEmbeddings
import os
AZURE_OPENAI_API_KEY = ""
os.environ["OPENAI_API_KEY"] = AZURE_OPENAI_API_KEY
loader = TextLoader(r"Toward a Knowledge Graph of Cybersecurity Countermeasures.txt")
documents = loader.load()
text_splitter = CharacterTextSplitter(chunk_size=1000, chunk_overlap=0)
chunks = text_splitter.split_documents(documents)
# create the open-source embedding function
embedding_function = SentenceTransformerEmbeddings(model_name="all-mpnet-base-v2")
current_directory = os.path.dirname("__file__")
# load it into Chroma and save it to disk
db = Chroma.from_documents(chunks, embedding_function, collection_name="groups_collection",
persist_directory=r"\rag_with_agent_chroma_db")
retriever = db.as_retriever(search_kwargs={"k": 5})
tool = create_retriever_tool(
retriever,
"search_state_of_union",
"Searches and returns documents regarding the state-of-the-union.",
)
tools = [tool]
llm = AzureChatOpenAI(
deployment_name='gtp35turbo',
model_name='gpt-35-turbo',
openai_api_key=AZURE_OPENAI_API_KEY,
openai_api_version='2023-03-15-preview',
openai_api_base='https://azureft.openai.azure.com/',
openai_api_type='azure',
streaming=True,
verbose=True
)
agent_executor = create_conversational_retrieval_agent(llm, tools, verbose=True, remember_intermediate_steps=True,
memory_key="chat_history")
result = agent_executor({"input": "hi, im bob"})
print(result["output"])
代码解读问题解决：
I attempted different approaches among the suggested solutions mentioned earlier. However, they failed to achieve the desired results. Even so, I attempted to lower its performance and upgraded The latest version of the LLM.
我对上述提到的各种方法进行了考察和探讨。然而，在实际操作中均未能达到预期效果。为了进一步优化性能, 我特意对LLM的不同版本进行了测试.
Finally, lastly, the following code used to initialize the agent succeeded in working for me with the current version of llm.
最后一步，请注意下述关键代码通过LLM（大型语言模型）实现自我运作并完成相应的逻辑处理。
conversational_agent = initialize_agent(
agent=AgentType.CHAT_ZERO_SHOT_REACT_DESCRIPTION,
tools=tools,
llm=llm,
max_iterations=10,
handle_parsing_errors=True,
early_stopping_method="generate",
memory=memory,
verbose=True,
)
代码解读
全部评论 (0)
还没有任何评论哟~