But the embedding modelll...?

The azure embedding model?

Yes

Seems like it does
https://github.com/run-llama/llama_index/blob/4187950c7cc5798a5836915cd39ac4ee4ded2304/llama_index/embeddings/azure_openai.py#L45

Must be in a later version than I have installed

Will try updating

ahaha https://github.com/run-llama/llama_index/commit/dacd6f53760e7f3396844725aaa005faea7dfb2a

hahaha yeaaaa

@Logan M with predictor gone, is the replacement in the LLM a 1:1?

ok it's not, i see

its essentially 1:1 ? They have the same interface afkaik

But you need to pass in llm=llm in the service context, not llm_predictor=llm

just some minor changes in 1 or 2 lines, but i got it figured out

this is a much better design btw

i love it

predictor always seemed unnecessary

I agree!

wait until you see v0.10.0 lol

no more service context (well, deprecated anyways)

can't wait to refactor 😂

jokes aside it'll be great

We made an upgrade script 🙂

can'tw ait