Whenever I load a model via HuggingFaceLLM in llama-index and pass it to a FunctionCallingAgent() as the llm object, I get a error "Model name {llm.model} does not support function calling API. "
I am using llama 3.1 as the llm.
Am I doing something wrong or is the HuggingFaceLLM() class does not somehow exposes the llm.metadata.is_function_calling_model property?
The same gives true for OpenAI class of models.