Hello i m facing some difficulties using

Did you check whether your engine supports GPT-4? You can go to Azure portal and check whether it is available in your region or not.
Azure gives a playground to check your deployed model. Maybe you give it a try there.

You can also check it here as well: https://learn.microsoft.com/en-us/azure/ai-services/openai/concepts/models#gpt-4-models

i'm not used to work with azure but the openai service where the gpt4 model is deployed is located on EAST US 2 while the openai service where the gpt3.5 model is deployed is on EAST US do you think my problem can come from there ?

(thanks a lot for trying to help me btw)

I checked further it says GPT-4 is only available by request. This could be the case for you
https://learn.microsoft.com/en-us/azure/ai-services/openai/chatgpt-quickstart?tabs=command-line&pivots=programming-language-studio#prerequisites

i'm not the one who deployed the gpt4 model but since it's deploy doesn't that mean i have access to gpt 4 ? (or i misunderstood the meaning of your message)

No, You have to select the model from Azure portal. So it's like which model resource you would like to choose from your Azure OpenAI. You choose one and that becomes available for you.

you mean that ?

Correct, So since you have this. There is a playground in Azure portal. Just open that and take all the values from there like api_version. base_url, engine for GPT-4 associated model.

Here's the guide on Azure: https://gpt-index.readthedocs.io/en/stable/examples/llm/azure_openai.html

ok but my problem wasn't that my problem is that i have created a rest api that link gpt4 with some indexed documents of mine (thanks to text-embedding-ada-002) so i'm not using azureopenai.create chatcompletion or chatmessage so i'm using the service_context from llama_index but it doesn't seem to work with gpt4 ? is it bcs i don't have the text-embedding-ada-002 model in the same environnement than my gpt 4 ?

Hello sorry to still ask for your help but i'm still locked on this issue do you maybe have another place where they could answer me if you are too busy are someone else that could help me ? (thx for the help so far sry if I have been rude with my last comment it was not supposed to be. I'm french so my english isn't that good sry again)

Hey @Lygatt no it was not rude at all!

I'm tagging @Logan M here, but could you please re iterate the issue again if possible. Otherwise Logan would have to read the entire conversation.

yes of course (and thanks 😉 )

So my issue is that :

i have an local api that work with flask langchain and azure to index a document on a mongoserv then i have some api request to ask chatgpt3.5-turbo-32k about the indexed document

my issue is that when i want to change the chatgpt model to a gpt4 model i get this error message :

i already checked if my api_base / key was correct and it is and my models deployment name aswell

so this is actually an error for embeddings

ok so it's bcs my embedding model is not on the same environnement than my gpt4 ?

Yea... since usually azure works by modifying globals, I think that might be causing issues 🤔

ok i'll ask my boss tomorrow if i can add a new model i'm in apprenticeship so i don't really know what i'm alowed to do are not. Thanks for the help i'll give you feedback tomorrow ❤️

Ok so not having the text-embedding-ada-002 model with my gpt4 was the issues it's now working after a little cleanup of the code. Thx a lot for the help provided @WhiteFang_Jr @Logan M ❤️ i wish you a good day/night

Awesome! 💪