Will Dean

Hi @ilaria-giunti Thanks for raising Can you post the full traceback

Think I know what it is. Take a look at the mmm.model_config. It now uses the Prior class for each of the distributions. You will will have to define the...

Can you post the error message instead of the code 🙏 Also, please post the version of the pymc-marketing you are using

I am seeing that is is a stale notebook. Can you try the solution `mmm.model_config["intercept"] = Prior("HalfNormal", sigma=5)` and report if that works

Add to the top ```python from pymc_marketing.prior import Prior ```

> but it says not mmm is not defined > > mmm.model_config["intercept"] = Prior("HalfNormal", sigma=5)--------------------------------------------------------------------------- NameError Traceback (most recent call last) Cell In[9], line 1 ----> 1 mmm.model_config["intercept"] = Prior("HalfNormal",...

You would want to add that in the function where `mmm` is a local variable

Change the function from ```python # BEFORE def create_and_fit_MMM(data: pd.DataFrame, target: pd.Series) -> DelayedSaturatedMMM: mmm = DelayedSaturatedMMM( date_column="Weeks", channel_columns=COORDS["media"], control_columns=COORDS["control"], adstock_max_lag=10, time_varying_intercept=True, sampler_config={ "chains": 4, "draws": 4000, "tune": 1000, "nuts_sampler":...

That's a separate package that needs to be installed. Try install our conda env

With recent updates to `gh` CLI tool, I am no longer having an issue with this. Has this changed for anyone else?