XGBoost multilabel classification on large datasets

Hi, I am trying to implement XGBoost for multilabel classification using XGBoost Classifier and sklearn’s OneVsRestClassifier. I need to train this model on large dataset split across multiple files. I am aware of partial_fit option of OneVsRestClassifier but that gives error with XGBoost Classifier, also about passing xgb_model in XGBoostClassifier for incrementa training but how to do so with OneVsRestClassifier in the case of multilabel classification