hi, our team is trying to replicate the XGB model built by DR model through python/R xgb function for validation purpose. For some reasons, even though we are using the same data, same parameters to start with, we ended up with different performance. Is there anybody with similar experience and have some insights on this?
Thank you very much!