Regarding RAG-end2end

Is there any way we can input our own retriever and generator and accordingly finetune our rag model ?
@shamanez