Do you have a full traceback example?
Let me try to reproduce tmr.
@Logan M Traceback (most recent call last):
File "/opt/anaconda3/lib/python3.9/site-packages/flask/app.py", line 2077, in wsgi_app
response = self.full_dispatch_request()
File "/opt/anaconda3/lib/python3.9/site-packages/flask/app.py", line 1525, in full_dispatch_request
rv = self.handle_user_exception(e)
File "/opt/anaconda3/lib/python3.9/site-packages/flask/app.py", line 1523, in full_dispatch_request
rv = self.dispatch_request()
File "/opt/anaconda3/lib/python3.9/site-packages/flask/app.py", line 1509, in dispatch_request
return self.ensure_sync(self.view_functions[rule.endpoint])(**req.view_args)
File "/Documents/GitHub/chat3pt-python/app.py", line 376, in query_model_product
results = query_product(openai_api_key, app.config["MODEL_BUCKET"], model_json_url, history_messages, query_str, detected_language, chunk_size_limit)
File "/Documents/GitHub/chat3pt-python/query_product.py", line 98, in query_product
response = query_engine.query(query)
File "/opt/anaconda3/lib/python3.9/site-packages/llama_index/indices/query/base.py", line 20, in query
return self._query(str_or_query_bundle)
File "/opt/anaconda3/lib/python3.9/site-packages/llama_index/query_engine/retriever_query_engine.py", line 139, in _query
nodes = self._retriever.retrieve(query_bundle)
File "/opt/anaconda3/lib/python3.9/site-packages/llama_index/indices/base_retriever.py", line 21, in retrieve
return self._retrieve(str_or_query_bundle)
File "/opt/anaconda3/lib/python3.9/site-packages/llama_index/indices/document_summary/retrievers.py", line 81, in _retrieve
raw_choices, relevances = self._parse_choice_select_answer_fn(
File "/opt/anaconda3/lib/python3.9/site-packages/llama_index/indices/utils.py", line 100, in default_parse_choice_select_answer_fn
answer_num = int(line_tokens[0].split(":")[1].strip())
IndexError: list index out of range
Thank you for your support
@Logan M More background information is that I have a bunch of documents , around 10-50 documents and the documents are used to construct the doc summary index like below. Sometimes the index works well and the LLM is able to answer my question based on the documents, but the above error will occur occasionally.
llm = ChatOpenAI(temperature=0, model_name="gpt-3.5-turbo", max_tokens=1000)
llm_predictor = LLMPredictor(llm=llm)
service_context = ServiceContext.from_defaults(llm_predictor=llm_predictor, chunk_size_limit=chunk_size_limit, callback_manager=callback_manager)
response_synthesizer = ResponseSynthesizer.from_args(response_mode="tree_summarize", use_async=True, service_context=service_context)
SUMMARY_QUERY = (
"Give a concise summary of the document in a paragraph.\n"
"If the document is a JSON of a product, the summary must include the 'Id' key.\n"
)
index = GPTDocumentSummaryIndex.from_documents(documents, summary_query=SUMMARY_QUERY, service_context=service_context, response_synthesizer=response_synthesizer)
Is it because of no relavant nodes to choose?
The above error is likely because of either no relevant nodes, or the output was not formatted correctly 🤔
If that is the case, I guess I can only catch the error and tell the user the LLM cannot answer the question? Anyway, appreciate for the help !
