OpenAI

At a glance

The community member is facing an error while using the VectorStoreIndex.from_documents() function from the llama_index library. The error is a "NotFoundError" with code 484 and message "Resource not found". The community members have shared their code and setup, which includes using the AzureOpenAI language model and embeddings. They have tried various configurations, including using the engine and deployment names, but are still facing the issue.

The solution provided by another community member is to define the service context and set it as the global service context. This involves creating a ServiceContext object with the llm and embed_model, and then setting it as the global service context using set_global_service_context(). This approach seems to have resolved the issue for the community member.

PPrince Kumar

Hey guys
Using openai 1.6.1
While doing
Index= VectorStoreIndex.from_documents(documents)

Getting this error
NotFoundError: Error code: 484 ['error': ['code": "404', 'message': 'Resource not found")).

Same error in this also.
index = VectorStoreIndex.from_documents(documents, 11m=11m1, embed_model =embeddings, prompt helper=prompt_helper)

21 comments

WWhiteFang_Jr

Could you share your code if possible

ttheoxd

what ll are you using (llama_index.llms) and how did u set it up

PPrince Kumar

AzureOpenAI

PPrince Kumar

import os
from dotenv import load_dotenv
from PyPDF2 import PdfReader
from llama_index import VectorStoreIndex, SimpleDirectoryReader, PromptHelper
from llama_index.llms import AzureOpenAI
from openai import OpenAI
from llama_index.embeddings import AzureOpenAIEmbeddings
import openai

load_dotenv()

openai.api_key = "YOUR_OPENAI_API_KEY"
openai.api_base = "YOUR_OPENAI_API_BASE"
openai.api_version = "2023-07-01-preview"

llm1 = AzureOpenAI(azure_deployment="gpt-35-turbo", azure_endpoint=openai.api_base, api_key=openai.api_key, api_version="2023-07-01-preview")

embeddings = AzureOpenAIEmbeddings(deployment="text-embedding-ada-002",
model="text-embedding-ada-002",
openai_api_base=openai.api_base,
openai_api_type="azure",
openai_api_key=openai.api_key,
openai_api_version="2023-07-01-preview")

Define prompt helper

max_input_size = 3000
num_output = 256
chunk_size_limit = 1000
max_chunk_overlap = 20

prompt_helper = PromptHelper(context_window=500, num_output=num_output, chunk_size_limit=chunk_size_limit)

documents = SimpleDirectoryReader('../data/qna/').load_data()

index = VectorStoreIndex.from_documents(documents)

ttheoxd

Do you have an azure subscription? are you sure about your deployement name?

ttheoxd

ik i never name them the exact model name

PPrince Kumar

Yess

PPrince Kumar

Using faiss db I have done these things

PPrince Kumar

Now trying it by llama indec

ttheoxd

AZURE_KWARGS: dict = {
"api_key": AZURE_OPENAI_API_KEY,
"azure_endpoint": AZURE_API_BASE_URL,
# "api_type": AZURE_API_TYPE,
"api_version": AZURE_API_VERSION,
"reuse_client": False,
}

Used it like this and it worked
def gpt_35(self) -> AzureOpenAI:
"""Return the GPT_35 model."""
return AzureOpenAI(
engine="OneBotGPT351106",
model="gpt-3.5-turbo-1106",
temperature=self.GPT_TEMP,
additional_kwargs=self.GPT_KWARGS,
**shared_settings.AZURE_KWARGS,
)

ttheoxd

You're missing the engine no?

PPrince Kumar

I tried using engine also inside the AzureOpenAI still facing error

WWhiteFang_Jr

Try and check if your llm and embed model are working.

Plain Text

print(llm.complete("hi"))

print(embeddings.get_text_embedding("hi"))

PPrince Kumar

Is engine different from deployment name?

ttheoxd

no it is the same

PPrince Kumar

It's working

WWhiteFang_Jr

Where is your service_context defined?

PPrince Kumar

I didn't defined it here but I tried it using service context it won't worked last time

WWhiteFang_Jr

Try this once:

Plain Text

from llama_index import ServiceContext
from llama_index import set_global_service_context
service_context = ServiceContext.from_defaults(
    llm=llm,
    embed_model=embed_model)
set_global_service_context(service_context)

WWhiteFang_Jr

Do this after defining the llm and embedding

PPrince Kumar

Its working great man thanks

Add a reply

Find answers from the community

OpenAI

Define prompt helper