how would I use a custom trained model I have in Gemini?
my use case, I’d like to create a chatbot with WAHA, that people can ask questions that get populated based on data of a Google Spreadsheet or DB. So people should be able to ask, which tickets are still open, or, provide a list of all open tickets. Now, we’ve trained a Gemini model that does this, but Gemini doesn’t have access to the live data, so we need n8n to pass that data to Gemini, and Gemini to use the custom model, and spit it back to n8n which passes it to WAHA.
Did you work it out in the end? The best I think of for this would be to use a vector store to keep your current data in and pull it from there like a lot of the RAG examples. Would this sort of approach work for you?
