oaieval command failure

[significant-gravitas]

oaieval command failure

[samuelbutler]

This line of your output looks like this isn't being run in a virtual env
[2023-04-19 06:03:19,041] [registry.py:249] Loading registry from /opt/homebrew/lib/python3.9/site-packages/evals/registry/completion_fns

That means the eval scripts can't find the benchmarking.pth file that you should have created in /Users/ejohnson/src/Auto-GPT-Benchmarks/venv/lib/python3.9/site-packages/benchmarking.pth after you created and activated a Python virtual environment named venv.

The contents of that benchmarking.pth file should be /Users/ejohnson/src/Auto-GPT-Benchmarks/ for your system.

That should resolve this error.

[johnseth97]

Hmm, that is what the contents of benchmarking.pth happens to be, but it's throwing the same error.

[samuelbutler]

even in a virtual environment? can you post your error output again while running this in a venv?

I had this same error and was able to resolve it which is what makes me think this isn't a bug.

[johnseth97]

After running source venv/bin/activate.fish, my fish prompt changes to show that I am in fact in a venv, but I still receive the same output.

I am running on a M1 MacBook Pro, but I really don't think that should make much of a difference.

[dschonholtz]

@desojo and @samuelbutler These guys should be resolved now.
Thanks for taking an initial crack at it when it was still really, really unstable.

[dschonholtz]

I am closing this, as this is no longer how we run benchmarks