MAD.py

Mad.py is a prototypical Python library for running Multi-Agent Debates (MADs).

Installation

Check out this repository and run pip install . or equivalent (e.g., for Conda).

Running a Basic MAD

Create an .env file in the examples folder. Specify variables as follows, for example:

MISTRAL_API_KEY = "<your_mistral_key_here>"
MODEL_PROVIDER = "mistral"
MODEL = "mistral-medium-latest"
LOG_LEVEL = "INFO"
SEMANTICS = "DFQuAD_model"
WAIT_FOR_RATELIMIT = 3

Then, run, e.g.:

python examples/basic_mad.py -t="['Kermit the Frog should be the next president of the USA.']"

Or:

python examples/basic_mad.py -t="['Frodo should be the next president of the US.', 'Bilbo should be the next president of the US.']"

Testing

Testing is done with pytest. Install test dependencies with pip install -e '.[dev]'. Then, run `pytest´.

Contributing

We welcome contributions, prepared with care and supported by tests.

Authors

• Timotheus Kampik - @TimKam

• Filip Naudot - @filipnaudot