You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This request is to add GPU support for prediction with leaf ids.
Our use case is to train a GBM model using XGBoost, then feed its leaf ids into another model. We find that running .predict with pred_leaf=True over our dataset is much slower than training hundreds of trees (on GPU) on the same dataset. I suspect this is because while training is GPU-accelerated (which is awesome!), prediction is not (0).
The text was updated successfully, but these errors were encountered:
This request is to add GPU support for prediction with leaf ids.
Our use case is to train a GBM model using XGBoost, then feed its leaf ids into another model. We find that running .predict with pred_leaf=True over our dataset is much slower than training hundreds of trees (on GPU) on the same dataset. I suspect this is because while training is GPU-accelerated (which is awesome!), prediction is not (0).
The text was updated successfully, but these errors were encountered: