Azure OpenAI Support?

[HRcode]

Unless I'm doing something wrong, it looks like connecting to an OpenAI model through Azure is not supported yet. I've tried to add my deployed Azure model name as a parameter called "deployment_id" (as that works in the Python version of LangChain) but I'm getting an "Incorrect API key provided" error in this version. I also tried "deploymentId", "deploymentID", and "engine" as parameter names.

Are there any plans to support Azure versions of OpenAI models?

[holwerda]

Was wondering about this too. Started taking a look at how to extend it.

[aaronrogers]

Is it possible to create a custom LLM to pass into LLMChain? If so, that'd help with this.

I opened a question about this and am hoping for an answer.

[nfcampos]

@aaronrogers I've now replied to you in the other thread. As for Azure OpenAI support, looks like this is an open issue on the openai-node client, see openai/openai-node#53

[yehudacohen]

@aaronrogers did you ever build a custom AzureOpenAI model? If so, can you share it as a gist or npm package?

[aaronrogers]

@yehudacohen, yeah, it's a bit slapped together, but let me see what I can do.

[aaronrogers]

@yehudacohen, it's not pretty, but it works. Your mileage may vary.
https://gist.github.com/aaronrogers/b6bb28e0e2299e5b96b168e8607a1cc2

[yehudacohen]

Thanks @aaronrogers , you're a hero!

[HRcode]

Thanks Aaron! Closing this for now. At some point it would be great for this to make it into the official langchainjs package, but I think this provides an onramp for Azure users to build a "custom" LLM for their langchain workflow.

[aaronrogers]

@HRcode, I agree, it would be nice to add. There's also some additional work that could be done to make the process of creating custom LLMs easier. There were a number of things I still felt like I needed to use from Langchain that weren't exposed, so I had to copy them over.

[thobhanifreddy]

Ideally, this should not be closed as it is not supported officially.