AttributeError: 'ColumnTransformer' object has no attribute '_name_to_fitted_passthrough' however versions of sklearn are the same

Question

model_loaded = joblib.load(path + model)
print(model_loaded.__getstate__()['_sklearn_version'])
res = model_loaded.predict(binance)

1.2.1

479 for _, name, transform in self._iter(with_final=False):

--> 480 Xt = transform.transform(Xt)

AttributeError: 'ColumnTransformer' object has no attribute '_name_to_fitted_passthrough'

print('The scikit-learn version is {}.'.format(sklearn.__version__))

The scikit-learn version is 1.2.1.

Before today all was fine

Answer 1

Actually, the method to check the pkl training version is not correct. I've enabeled warnings and it showed me that I was using 1.1.1 so downgrading solved the problem