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
Hello I was training my classifier using XGBoost with the following code on GCP and eventually I built my own server and now I am having difficulty. On GCP I used to see the message 'backend: using loky backend ...' but not on my own server which makes me think if this code is really taking advantage of parallelism. The server I built is far superior to what I was using on GCP. Below is some information about my code and the specs of the server. I wonder is there is something wrong with my environment.
If I use n_jobs=-1 inside the RandomizedSearchCV() code then I start getting the error below. However if I use n_jobs=-1 inside XGBClassifier() then things work albeit it is slow which makes me think that I am not taking advantage of parallelism. I also do not see the 'Backend: using loky backend ...' message either.
warnings.warn("Estimator fit failed. The score on this train-test"
/home/michael/.local/lib/python3.8/site-packages/sklearn/model_selection/_validation.py:610: FitFailedWarning: Estimator fit failed. The score on this train-test partition for these parameters will be set to nan. Details:
Traceback (most recent call last):
File "/home/michael/.local/lib/python3.8/site-packages/sklearn/model_selection/_validation.py", line 593, in _fit_and_score
estimator.fit(X_train, y_train, **fit_params)
File "/home/michael/.local/lib/python3.8/site-packages/xgboost/core.py", line 422, in inner_f
return f(**kwargs)
File "/home/michael/.local/lib/python3.8/site-packages/xgboost/sklearn.py", line 913, in fit
self._Booster = train(xgb_options, train_dmatrix,
File "/home/michael/.local/lib/python3.8/site-packages/xgboost/training.py", line 206, in train
bst = _train_internal(params, dtrain,
File "/home/michael/.local/lib/python3.8/site-packages/xgboost/training.py", line 101, in _train_internal
bst.update(dtrain, i, obj)
File "/home/michael/.local/lib/python3.8/site-packages/xgboost/core.py", line 1280, in update
_check_call(_LIB.XGBoosterUpdateOneIter(self.handle,
File "/home/michael/.local/lib/python3.8/site-packages/xgboost/core.py", line 189, in _check_call
raise XGBoostError(py_str(_LIB.XGBGetLastError()))
xgboost.core.XGBoostError: [17:24:00] ../src/c_api/../data/../common/common.h:44: ../src/c_api/../data/../common/device_helpers.cuh: 128: out of memory
Stack trace:
[bt] (0) /home/michael/.local/lib/python3.8/site-packages/xgboost/lib/libxgboost.so(+0x8d264) [0x7fccd8733264]
[bt] (1) /home/michael/.local/lib/python3.8/site-packages/xgboost/lib/libxgboost.so(+0x2fc730) [0x7fccd89a2730]
[bt] (2) /home/michael/.local/lib/python3.8/site-packages/xgboost/lib/libxgboost.so(+0x2fcde9) [0x7fccd89a2de9]
[bt] (3) /home/michael/.local/lib/python3.8/site-packages/xgboost/lib/libxgboost.so(+0x32da14) [0x7fccd89d3a14]
[bt] (4) /home/michael/.local/lib/python3.8/site-packages/xgboost/lib/libxgboost.so(+0x339bc5) [0x7fccd89dfbc5]
[bt] (5) /home/michael/.local/lib/python3.8/site-packages/xgboost/lib/libxgboost.so(+0x33aa2b) [0x7fccd89e0a2b]
[bt] (6) /home/michael/.local/lib/python3.8/site-packages/xgboost/lib/libxgboost.so(+0x431dc3) [0x7fccd8ad7dc3]
[bt] (7) /home/michael/.local/lib/python3.8/site-packages/xgboost/lib/libxgboost.so(+0x4384bd) [0x7fccd8ade4bd]
[bt] (8) /home/michael/.local/lib/python3.8/site-packages/xgboost/lib/libxgboost.so(+0x17255b) [0x7fccd881855b]
Hello I was training my classifier using XGBoost with the following code on GCP and eventually I built my own server and now I am having difficulty. On GCP I used to see the message 'backend: using loky backend ...' but not on my own server which makes me think if this code is really taking advantage of parallelism. The server I built is far superior to what I was using on GCP. Below is some information about my code and the specs of the server. I wonder is there is something wrong with my environment.
If I use n_jobs=-1 inside the
RandomizedSearchCV()
code then I start getting the error below. However if I use n_jobs=-1 insideXGBClassifier()
then things work albeit it is slow which makes me think that I am not taking advantage of parallelism. I also do not see the 'Backend: using loky backend ...' message either.MY HARDWARE:
The text was updated successfully, but these errors were encountered: