Continue training from a loaded model using BaggingClassifier

Question

I currently have the following imbalanced learn Pipeline set-up:

Pipeline(steps=[('sampling',
                 FunctionSampler(func=<function ...>)),
                ('classification',
                 BalancedBaggingClassifier(base_estimator=XGBClassifier(..., n_estimators=10))
               ])

using imblearn, sklearn and xgboost. I am saving the trained model using the joblib library and loading it up again in another script. My goal is then to train the pre-trained model from the loaded starting point.

How I would do this using just the XGBClassifier is something like this (see here):

model = model.fit(X_train, y_train, xgb_model=model.get_booster())

where model would be loaded using joblib.

However, in the boosting case, I essentially have n_estimators many estimators. I know how to access them individually, but how can I pass all of them to the pipeline fit function?