Problem with CuDNN initialization - Technical Help - DeepTalk - Deep Learning Community
Problem with CuDNN initialization
post by WinSurfer on Dec 10, 2020
Dear Technical help,
I have installed lambda-stack in my system (ubuntu) and installed Pycharm with the “inherit global site-packages” setting to use the lambda-stack environment.
However, when I try to run some demo Tensorflow jobs, I got the error of CuDNN initialization (figure attached). Do you have any suggestions for fixing this problem?
Many thanks!
post by jeremy on Dec 14, 2020
Try adding the following after importing Tensorflow:
for device in tf.config.experimental.list_physical_devices('GPU'):
tf.config.experimental.set_memory_growth(device, True)