You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
While training a network with more than 1 training epoch, after the first training epoch completes, the following error occurs (here I was using just 3 epochs):
"An error occurred while calling the 'BokehVisualizer' hook:
TypeError: range(0, 2) is not JSON serializable
I am running Python 3.4 and using Bokeh version 0.9.0, on a Windows 10 machine (and my Bokeh setup works as I am able to run it successfully with other packages).
Additional information:
When debugging, the error is originating from the anaconda3\lib\json\encoder.py file:
In this file, there is a method called "default" containing a try-except that handles arbitrary iterators that was commented out. After I uncommented it out, the BokehVisualizer's type error no longer occurs. However, what happens now is that after the first training epoch, instead of throwing the type error, ALL plot information (axes, title, and legend) disappears and all that is left is a blank plot that does not display any information.
At line 815 of brainstorm\hooks.py there is a range() being passed in to the json\encoder file. Would this behave differently between Python2 and Python3?
The text was updated successfully, but these errors were encountered:
While training a network with more than 1 training epoch, after the first training epoch completes, the following error occurs (here I was using just 3 epochs):
"An error occurred while calling the 'BokehVisualizer' hook:
TypeError: range(0, 2) is not JSON serializable
I am running Python 3.4 and using Bokeh version 0.9.0, on a Windows 10 machine (and my Bokeh setup works as I am able to run it successfully with other packages).
Additional information:
When debugging, the error is originating from the anaconda3\lib\json\encoder.py file:
In this file, there is a method called "default" containing a try-except that handles arbitrary iterators that was commented out. After I uncommented it out, the BokehVisualizer's type error no longer occurs. However, what happens now is that after the first training epoch, instead of throwing the type error, ALL plot information (axes, title, and legend) disappears and all that is left is a blank plot that does not display any information.
At line 815 of brainstorm\hooks.py there is a range() being passed in to the json\encoder file. Would this behave differently between Python2 and Python3?
The text was updated successfully, but these errors were encountered: