Give Feedback To Autogpt

[jaylewin]

Does anyone know how to give feedback to autogpt in the langchain setup file for web assistant?

https://python.langchain.com/en/latest/use_cases/autonomous_agents/marathon_times.html

How can I add a cell in a jupyter file for adding feedback to an autogpt agent each time it loops?

Here is the code I have that will print the agent results, but once it starts I can't figure out where to input my feedback for it's loop.

---

tools = [
web_search,
WriteFileTool(root_dir="./data"),
ReadFileTool(root_dir="./data"),
process_csv,
query_website_tool,
HumanInputRun(), # Activate if you want the permit asking for help from the human
]

agent = AutoGPT.from_llm_and_tools(
ai_name="Tim",
ai_role="Assistant ",
tools=tools,
llm=llm,
memory=vectorstore.as_retriever(search_kwargs={"k": 8}),
human_in_the_loop=True, # Set to True if you want to add feedback at each step.
)
print(agent) # Add this line to check if the agent variable has been initialized correctly
while True:
output = agent.run(["help me to a great thing."])
print(output)
feedback = input("I want to see something else")
agent.add_memory(feedback)
__

Thanks