If you have a problem, that python program hangs for about a minute during running of TensorFlow program on GPU, probably this solution is for you.

1. The problem

Let’s suppose that you have a hello-world program like below.

from __future__ import print_function

import tensorflow as tf

a = tf.constant("Hello World")
session = tf.Session()
output = session.run(a)

print('Output:' + str(output))

While you run it, you can notice that program hangs for about 60 seconds displaying the following log:

D:\__BC\workspace\tensorflow-tutorial\01-hello-world\venv1\Scripts\python.exe D:/__BC/workspace/tensorflow-tutorial/01-hello-world/hello-world.py
2019-05-03 20:35:28.199069: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2
2019-05-03 20:35:28.227552: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library nvcuda.dll
2019-05-03 20:35:28.925366: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1467] Found device 0 with properties: 
name: GeForce GTX 960M major: 5 minor: 0 memoryClockRate(GHz): 1.176
pciBusID: 0000:01:00.0
totalMemory: 4.00GiB freeMemory: 3.34GiB
2019-05-03 20:35:28.925960: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1546] Adding visible gpu devices: 0

2. The solution

Add below lines at the beginning of your python code:

import os
os.environ['TF_CPP_MIN_LOG_LEVEL']='2'

In my previous example it may looks like below:

In some cases it is required to restart IDE.

from __future__ import print_function

import tensorflow as tf
import os
os.environ['TF_CPP_MIN_LOG_LEVEL']='2'

a = tf.constant("Hello World")
session = tf.Session()
output = session.run(a)

print('Output:' + str(output))