Hi I'm trying to execute on mac machine this example workflow with path: 04_Analytics/14_Deep_Learning/02_Keras/04_Cats_and_Dogs/02_Train_simple_CNN
So, basically I have installed python 3.5 with all Keras requirements with:
conda create -y -n py35_knime python=3.5 pandas jedi keras
I have also activated that environment. After that I have put the exec path for python 3.5 environment in preferences > KNIME > python, and set as "default".
So, after successfully execute the previous workflow "preprocessing images" I encounter several errors in this one when I execute the first node DL Python Network Creator:
ERROR PythonKernel socket.gaierror: [Errno 8] nodename nor servname provided, or not known
ERROR DL Python Network Creator 0:222 Execute failed: Could not start python kernel. Cause: The connection attempt timed out. Please consider increasing the socket timeout using the VM option '-Dknime.python.connecttimeout=<value-in-ms>'.
Any suggestions?.. I'm just trying to execute the workflow in the current state.
there seems to be a communication problem between KNIME and the Python process where Keras is run.
Could you have a look at the host file of your computer - you can find it under /etc/hosts - and check if there's an entry "127.0.0.1 localhost" or the like?
You can navigate to the file using Finder (or by looking for "/etc" in Spotlight Search) and then open it using Text Edit.
If you're comfortable with the terminal, you can also simply open up a terminal window and type in
##
# Host Database
#
# localhost is used to configure the loopback interface
# when the system is booting. Do not change this entry.
##
127.0.0.1 localhost
255.255.255.255 broadcasthost
::1 localhost