CODEWITHSUNDEEP
pythongpuxgboost
EoinMcL
EoinMcL

Reputation: 85

Getting Error " Check failed: gpu_predictor_: " when trying to use XGBoost with GPU in Python

I am trying to use XGBoost in Python getting the error 

    xgboost.core.XGBoostError: [09:26:06] c:\users\administrator\workspace\xgboost-win64_release_1.0.0\src\gbm\gbtree.h:308: Check failed: gpu_predictor_:

I am using Sklearn and construction the model like

    model = XGBRegressor(learning_rate = 0.02,
                 n_estimators = 1000,
                    max_depth = 12,
                    tree_method='gpu_hist',
                    objective = 'reg:logistic',
                    nthread = -1,
                    silent = False)

It works with "hist" as tree_method on cpu.

Upvotes: 1

Views: 1988

Answers (1)

akra1
akra1

Reputation: 185

Maybe your GPU is not supporting CUDA which is required according to the XG Boost website:

The GPU algorithms in XGBoost require a graphics card with compute capability 3.5 or higher, with CUDA toolkits 10.0 or later

At least, this was my problem. If you want to check it, you can use this list.

Sadly, I couldn't find a workaround for this problem.

Upvotes: 2

Related Questions