We need a way to deterministically identify if an FLM model is a chat-capable LLM model. Right now, AFAIK, FLM only has labels for things like embedding and transcription models to indicate they are not chat models, but no explicit chat label.
We are adding the chat label to Lemonade here: lemonade-sdk/lemonade#3099
This will make client development much easier.
We need a way to deterministically identify if an FLM model is a chat-capable LLM model. Right now, AFAIK, FLM only has labels for things like embedding and transcription models to indicate they are not chat models, but no explicit chat label.
We are adding the chat label to Lemonade here: lemonade-sdk/lemonade#3099
This will make client development much easier.