How to hide warnings from xgboost library in jupyter?

Question

Not working:

import warnings
warnings.filterwarnings('ignore')

The warning I get:

[14:24:45] WARNING: C:/Jenkins/workspace/xgboost-win64_release_0.90/src/objective/regression_obj.cu:152: reg:linear is now deprecated in favor of reg:squarederror.

It clutters my output in cell.

found it, XGBRegressor(objective ='reg:squarederror') – kaban Oct 10 '19 at 08:39 — kaban, Oct 10 '19 at 08:39
please put your comment as an answer to help others – Amirkhm Dec 06 '19 at 12:19 — Amirkhm, Dec 06 '19 at 12:19

score 30 · Answer 1 · edited Jun 10 '20 at 11:53

30

Change the verbosity parameter verbosity = 0 in the model definition. The values it can take are: 0 - "silent", 1 - "warning", 2 - "info", 3 - "debug"

xgboost = xgb.XGBRegressor(objective ='reg:linear', verbosity = 0, random_state=42)

XGBoost Documentation

edited Jun 10 '20 at 11:53

Adrian Mole

49,934
160
51
83

answered Jun 10 '20 at 11:40

ASHISH JETHANI

401
4
4

7

This is actually the only answer that works (in 2020). silent=True does not work, neither does warnings.filterwarnings('ignore') if you are about to try that. Not sure if this has to do with XGBRegressor not actually classifying its warning messages correctly? Nonetheless verbosity = 0 is what kills of the 100's of endless warnings that occur during a gridsearch for example. – Christoffer Oct 15 '20 at 08:13
VSCode, Jupyter, Python 3.8, Linux Mint - works, but doesn't quiet UserWarnings. – Thom Ives Sep 16 '21 at 23:08

score 5 · Answer 2 · edited Dec 17 '19 at 15:22

5

Just add silent = True in the model's definition:

xgboost = xgb.XGBRegressor(random_state=42,silent=True)

edited Dec 17 '19 at 15:22

Adrian Mole

49,934
160
51
83

answered Dec 17 '19 at 15:00

Aleksandra Afanaseva

51
1
3

2

This added even more warnings! – Bex T. Apr 08 '21 at 09:49
Use both as @darkstar-dream verbose=0 and silent=True – Rajdeep Borgohain Mar 03 '22 at 16:16

score 1 · Answer 3 · answered Sep 18 '21 at 04:04

1

I had the same issue, and adding both verbose=0 and silent=True helped me

xgboost = xgb.XGBRegressor(objective ='reg:linear', verbosity = 0, silent=True, random_state=42)

answered Sep 18 '21 at 04:04

Darkstar Dream

1,649
1
12
23

How to hide warnings from xgboost library in jupyter?

3 Answers3