You can also implement customized LLMs or VLMs as agents to test out the long-chain reasoning ability of the models. The agents should be implemented inagent.pyunderagent/folder by searching foryour_customized_modeltag. Remeber to initializa a new agent instance and modify the respective config...