Live feedback with python scripting

[l.thomas]

Hi,
I was wondering since there are tools like tensorboard to monitor machine learning training in the browser.
Would it be possible with the python nodes to have any generic feedback (print statement…) similarly forwarded to the browser or to some external tool ?
Currently there is know way to get some idea of the advancement, so such workaround could be a solution.

[christian.birkhold]

Hi @l.thomas,

For the “Keras Network Learner” you can already monitor the progress of the training process. Not yet in the web, though.

Cheers,

Christian

[l.thomas]

Hi @christian.dietz
I know about this one, but I was thinking for generic python node that would be great too !

I tried to redirect my print statement to knime.log, I can see my statements in the file but they are not displayed in the console panel of KNIME.
Do you think I could find a way there ?

The python code :

log = r"C:\Users\Laurent\knime-workspace\.metadata\knime\knime.log"

with open(log, "a") as f:
    print("2019-10-22 12:07:01,956 : WARN  : KNIME-Worker-14 :  : Node : Python Source : 0:6 : Hello KNIME!", file=f)

# Create am empty datatable to prevent execution error
import pandas as pd
output_table = pd.DataFrame()

[nemad]

Hello @l.thomas,

you can display the standard output of Python in the console panel by setting the “Console View Log Level” to INFO (File->Preferences->KNIME->KNIME GUI).

Cheers,

Adrian

[l.thomas]

AMAZING !!!