[Python] How to use set_config() method of sktime objects

I want to perform parallel computing on my ShapeletTransform. According to the documentation, I need to set the backendarallel flag to 'loky' to activate the parallel method of the class. However, the set_config method accepts "**config_dict" as parameter.

Here's what I tried:

y_transformed = y

y_transformed = pd.DataFrame(y.apply(lambda row: pd.Series(row), axis=1))

transformer = ShapeletTransform(verbose=1)
dict_config = {
"backendarallel": "default=loky"
}
transformer.set_config(dict_config)

transformer.get_config()

transformer.fit(X_3d, y_transformed)


Here's the error:

TypeError Traceback (most recent call last)
\<ipython-input-96-2327096163\> in \<cell line: 0\>()
12 "backendarallel": "default=loky"
13 }
\---\> 14 transformer.set_config(dict_config)
15 # Get the current configuration
16 transformer.get_config()


TypeError: BaseObject.set_config() takes 1 positional argument but 2 were given

Continue reading...