Find answers from the community

Updated 3 months ago

Hi, after updating LlamaIndex to the

Hi, after updating LlamaIndex to the latest I see the exception "'openai' has no attribute 'error'" fired here "site-packages\langchain\chat_models\openai.py", line 77" but not sure how to fix it. Thanks!

20 comments

LLogan M

Try pip uninstall llama-index openai and then pip install llama-index

SSeaCat

Thanks, I will try!

SSeaCat

@Logan M Tried but it still installs openai 1.2.0 and this version already doesn't have error in their init.py file

LLogan M

right, and the latest version of llama-index doesn't import from error

LLogan M

the traceback is from langchain

LLogan M

lol rip

LLogan M

is there a reason you are using langchain here?

LLogan M

we have our own LLM objects

SSeaCat

No reason at all, I thought it's from you 🤔

LLogan M

site-packages\langchain\chat_models\openai.py

LLogan M

langchain 🙂

LLogan M

Plain Text

from llama_index.llms import OpenAI

llm = OpenAI(model="gpt-4", temperature=0.1)

SSeaCat

Not sure what is that

LLogan M

I pointed out the error was coming from langchains files, and then gave an example of how to use our LLM instead 👍

SSeaCat

Ahhh okay, I see. In my code I can see these imports:

Plain Text

from langchain.prompts.chat import (
    AIMessagePromptTemplate,
    ChatPromptTemplate,
    HumanMessagePromptTemplate
)

from langchain.chains.prompt_selector import ConditionalPromptSelector, is_chat_model
from langchain.chat_models import ChatOpenAI

I copied all of them some time ago from examples on your website...

LLogan M

times change I suppose 🙂

SSeaCat

So, do you think I can find all these classes/functions in LlamaIndex without using langchain?

LLogan M

Depends what you are doing with all those. At the very least, you can probably replace ChatOpenAI with our OpenAI module and avoid the error I think?

SSeaCat

Okay, gotcha! Thanks for your help!!

Add a reply