Find answers from the community

Updated 3 months ago

I am having issues with the embed model

ttomoe

I am having issues with the embed model from azure not loading

Attachment

6 comments

WWhiteFang_Jr

LlamaIndex defaults to OpenAI embedding and llm model. If you want to use diff ones you'll have to pass them in the methods or define them globally using Settings.

Plain Text

from llama_index.core import Settings
Settings.llm= YOur Azure llm
Settings.embed_model = Your embed_model instance here

This will now use your llm and embedding model everywhere

ttomoe

so i just have to change to

Settings.embed_model = AzureOpenAIEmbedding(
model="text-embedding-3-large",
deployment_name="text-embedding-3-large",
azure_endpoint=AOAI_API_BASE,
api_key=AOAI_API_KEY,
api_version=EMB_API_VERSION,
)

and it will work again?

ttomoe

But I was following a close example, why would this not work, previously it worked flawlessly

Attachment

WWhiteFang_Jr

If the example contains an azure based llm and embed model then they must be passing it in every method.

WWhiteFang_Jr

Yes you gotta do this for llm as well

ttomoe

Thank you, issue fixed

Add a reply