Bambi multiple cores in jupyter notebook on remote server, no sampling

Questions
1

Hi all,

I tried fitting a model implemented with bambi on a jupyter notebook run on a remote server (using the --no-browser argument). When I specify a number of cores bigger than 1 (when calling model.fit()), the progress bar does not move at all for a long time (I gave up after some minutes so I don’t know if it would start moving at some point). However, with the model implemented directly in pymc3, this works (i.e. specifying cores=8 in the pm.sample() command). Has anyone experienced this as well and possibly knows how to solve it?

Thanks!

2

Hi @leomein, do you have a reproducible example of the model you’re trying to fit so I can try to reproduce it on my side?

Thanks!

3

Hi, thanks for your quick reply! It seems like this was a temporary problem of the computer cluster I was on and it looks like it’s working now. Sorry about this!

1 Like