Using Multiple OpenAI Clients for Different Models
I have this situation where I want to use two OpenAIClients.
For embeddings, I want to use the original OpenAI client, but for the language models, I want to use OpenRouter.
So whenever I provide the embedding model, I also need to provide the right client otherwise it's going to go to the wrong API (for example OpenRouter doesn't support embedding models).
So I created a LayerMap but I'm not sure there's much benefit.
How should I do this?
For embeddings, I want to use the original OpenAI client, but for the language models, I want to use OpenRouter.
So whenever I provide the embedding model, I also need to provide the right client otherwise it's going to go to the wrong API (for example OpenRouter doesn't support embedding models).
So I created a LayerMap but I'm not sure there's much benefit.
How should I do this?
