Suppression of deprecation warnings in Tensorflow Python
Hey guys, this article will be focussing on the deprecation warnings in TensorFlow as well as the logging information.
As with the new update of TensorFlow, from 1.x to 2.x, the deprecation warnings have been removed by default but still, some logging information is provided. As fas far, TF 1.x is concerned, we still need to work upon the deprecation warnings and informational loggings.
Implementation for suppression of deprecation in Tensorflow:-
For TF 1.x, the following code can be used to remove logging information:-
For TF 2.x:-
Or we can change the contrib warnings to none to stop prevent warnings to get printed:-
import tensorflow as tf if type(tf.contrib) != type(tf): tf.contrib._warning = None
Else we can add flexibility to our work by using the log_level codes to choose what to show and what not to:-
import os import tensorflow as tf os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'
where:- 0 = all messages are logged. 1= INFO logs are removed. 2 = INFO with WARNINGS is removed. 3= ALL messages are removed.
Now its time to remove the deprecated warnings for TF 1.x as 2.x already handles it without displaying. So this code remove all the deprecated warnings:-
from tensorflow.python.util import deprecation deprecation._PRINT_DEPRECATION_WARNINGS = False
Or if we want future warnings to be suppressed too with the current deprecated warnings, the following can be used:-
import warnings warnings.filterwarnings('ignore', category=DeprecationWarning) warnings.filterwarnings('ignore', category=FutureWarning)
I hope you liked this article, and do check our other posts.