Find answers from the community

Updated 2 months ago

Leveraging Agent Runner with Gemini-Pro Model

At a glance

The community member is trying to use the Agent Runner tool to develop a framework with the Gemini-Pro model instead of OpenAI. However, they are receiving a response stating that the available tools lack the desired functionality. The community members discuss that this is not an error, but rather the language model's inability to solve the query with the current tools. They suggest modifying the tool names/descriptions or the system prompt to potentially resolve the issue. The community members also discuss that Agent Runner is preferably used with OpenAI models, and that Gemini Pro may not be suitable for use with Agent Runner. The community member is considering using Llama 3.2 from Meta as an alternative to work with Llama Index to develop a RAG framework, as they are working with sensitive data and ChatGPT would not be accepted for legal reasons.

Hi @Logan M
I have a question, I am trying to use Agent runner to develop a framework with gemini-pro model instead of Open AI
When I do it and do the agent.query and put in my question
I get this response:
I am sorry I cannot fulfill this request. The available tools lack the desired functionality.

What's this error about? What are the models that can leverage agent runner outside of OpenAI? Can Llama 3.2 from Meta do the same ?
L
h
5 comments
Its not an error, the LLM simply thinks that the query cannot be solved by any of the available tools
Probably modifying the tool names/descriptions or the system prompt would help
Let me give it a try, I researched a bit via Claude and Perplexity, the answers I got was Agent Runner is preferably to use with OpenAI model but not other models like Gemini Pro, is that correct?
Thanks a lot for helping @Logan M
Its pretty much the same yea πŸ‘
Do you mean Gemini Pro is not suitable to work with Agent Runner?
The reason I am using Gemini Pro instead of OpenAI is because the city I am living in , OpenAI do not provide services
I am using Gemini Pro as a testing ground with an ultimate goal to adopt Llama 3.2 from Meta to work with Llama Index to develop a RAG framework.
I am constantly working with sensitive data and ChatGPT would not be accepted for the legal reason.
Add a reply
Sign up and join the conversation on Discord