I am using an open source llm with an agent all through llamaindex. it says the agent is using the tools, however, it's actullly not using it and at the end of each response from (agent.stream_chat) it says Agent: None.
Anyone knows what might be going on and why it's not using the tool and why Agent: None.?
Thank you so much for your response @WhiteFang_Jr . My code follows the exact same setup (first link), but still the same issue and I get Agent: None. The only diference is that if I use agent.chat() instead of stream_chat() method, it doesn't put an Agent: None at the end of the response, but is still not using the tool. It writes down all the thought process and says that i'll use the tool, but it doesn't actually do it. (but if I use openai model and the openai agent, it works as expected, however, I need to use open source stuff for privacy reasons).
As for the second link that you shared, I was using "llama2", "mistral", and "gemma" models where all of them pretty much failed to use that single tool that I defined for it (which is just a python function). I wonder if you have any suggestion on an open source model that does well with agents (I need to use certain functions as tools) for a chatbot?
I appreciate your time and response in advance, thank you.
