Find answers from the community

Updated last year

You can create ChatOpenAI object and

You can create ChatOpenAI object and pass it to the service_context like this

Plain Text

from langchain.chat_models import ChatOpenAI
llm = LLMPredictor(llm=ChatOpenAI(openai_api_key=OPENAI_API_KEY,temperature=0, max_tokens=1024, model_name="gpt-3.5-turbo"))

Then pass this into

Plain Text

service_context = ServiceContext.from_defaults(chunk_size_limit=512, llm=llm)

6 comments

LLogan M

I think this is slightly incorrect 👀

LLogan M

Plain Text

from llama_index.llms import OpenAI

llm = OpenAI(model="gpt-3.5-turbo", temperature=0)
service_context = ServiceContext.from_defaults(llm=llm)

LLogan M

@keychron you may need to share your code again, not sure why it's using davinci baded on the last time I saw it 👀

WWhiteFang_Jr

Langchain ChatOpenAI is not being used now 😅?

LLogan M

We have our own LLM abstractions now 💪 They work a little better across llama-index features (query engines, chat engines, agents) https://gpt-index.readthedocs.io/en/latest/core_modules/model_modules/llms/modules.html

kkeychron

@Logan M full code here

Add a reply