The community members are discussing how to use the Xinference API for language model embeddings in the llama-index library. They consider extending the OpenAI embedding metadata or waiting for a dedicated llama-index-embeddings-openai-like feature. Some suggestions include creating a custom embedding class, using the ollama library, or modifying the OpenAIEmbedding class to work with the Xinference API. The community members also discuss the complexity of the llama-index codebase and how to best approach the integration.
wrt the code, i guess it's in some hybrid state and the intend is to make them all separate repos? anyway, atm it looks odd but i understand why this is better