python-agents-tutorial/2_local

At a glance

The community member is having trouble running a simple example script on their M1 Mac with 32GB of RAM. They get an exit code of 0 and no response, even when trying a smaller model. The community members suggest trying a different model like llama3.x, checking if the data folder is properly configured, and ingesting embeddings. They also recommend using a larger model like 8B parameters for better performance. While there is no explicitly marked answer, the community members provide several suggestions to help the original poster troubleshoot the issue.

Useful resources

mmarvanni

So Maybe I get this all wrong, but Im banging my head around a simple example for days now...

I work on a M1 mac with 32GB ram.

IM trying to get this example to run:

https://github.com/run-llama/python-agents-tutorial/blob/main/2_local_agent.py

But all I get is "Process finished with exit code 0"

I try to use a smaller model, but this als does not generate the response..

When I chat with the llm directy it does deliver output.

Are these example even ment to run on local machines?

12 comments

LLogan M

This works on my M2 + 32GB

Maybe try a different model? llama3.x is probably a better choice these days

LLogan M

I'm guessing you ran the script with python ./2_local_agent.py ?

mmarvanni

Yeah im using

Settings.llm = Ollama(model="llama3.2:1b")

and run it from pycharm, not sure if that matters..

mmarvanni

even this example does not work, so wonder if there is something on the machine that needs to be configured?

https://docs.llamaindex.ai/en/stable/getting_started/starter_example_local/

LLogan M

usually empty response is returned when no documents are retrieved 🤔 Does data/paul_graham point to a real folder with real text in it?

mmarvanni

yeah.... i created a clean project, that does run the example now....

mmarvanni

its like learning to programm for this first time.... 😦

Attachment

GGianluca

Empty response on my experience is when the documents are not ingested, try ollama pull mxbai-embed-large

GGianluca

And create a embedding with ollama

GGianluca

https://docs.llamaindex.ai/en/stable/examples/embeddings/ollama_embedding/

WWhiteFang_Jr

Hey if you check the following doc: https://docs.llamaindex.ai/en/stable/examples/agent/react_agent/#define-function-tools

The defined functions have docstring. Could you try adding it and run again?

Also 1b parameter model is small for performing function calling. I would suggest try 8b parameter once

mmarvanni

thanks, will try after dinner!

Add a reply

Find answers from the community

python-agents-tutorial/2_local_agent.py ...