search

AxeldeRomblay / MLBox

MLBox is a powerful Automated Machine Learning python library.
https://mlbox.readthedocs.io/en/latest/
Other
1.49k stars 274 forks source link

Hub connection request timed out #29

Closed ilyes495 closed 7 years ago

ilyes495 commented 7 years ago

I tried to run the code in the picture below, but I got the error saying TimeoutError: Hub connection request timed out. I'm using Python2.7 under Ubuntu 16.04

hub connection time out

Thanks for your help

AxeldeRomblay commented 7 years ago

Hi, Actually if you import mlbox.preprocessing, you have to wait approx 30 sec so that engines are ready... You can check in terminal that the following message appears : 'the engines appear to have started successfully'

Here is the doc for ipyparallel: https://ipyparallel.readthedocs.io/en/latest/process.html

At the moment, you have to kill your kernel or shutdown your VM, restart your code and wait 30 sec before running fit_transform

maskani-moh commented 7 years ago

@AxeldeRomblay

What about warning the user when instanciating a Drift_thresholder. (Do a print basically asking to wait) ?

ilyes495 commented 7 years ago

I restarted the kernel, run the imports and waited for more than 30 sec, when I ran the code, I got the same error

ilyes495 commented 7 years ago

@maskani-moh I don't get a warning when instantiating drift_thrsholder

maskani-moh commented 7 years ago

@ilyes495

I know that. What I was suggesting to the developers is to put a warning to let the user wait 30s, which is not the case right now.

AxeldeRomblay commented 7 years ago

@ilyes495. Did you get the warning when importing the lib ? What you can do: os.system("ipcluster stop --profile=home &") and then restart: os.system("ipcluster start --profile=home &") and then wait for the engines to be ready, etc... . Otherwise restart your computer ???

Anyway, we will get rid of ipyparallel for next releases as it is too complicated to use (it is helpful if you want to distribute on several machines...)

ilyes495 commented 7 years ago

I ran these commands and the problem was solved. jupyter serverextension enable --py ipyparallel jupyter nbextension install --py ipyparallel jupyter nbextension enable --py ipyparallel

ilyes495 commented 7 years ago

I got an error after that on preprocessing file line 148 "fichier.write(var + " = " + str(d) + '\n')"
saying: TypeError: unsupported operand type(s) for +: 'int' and 'str', could be that var is an int rather than a string?

ilyes495 commented 7 years ago

I solved the previous problem, by adding str(var).

AxeldeRomblay commented 7 years ago

thanks for reporting this issue :)