Closed denijane closed 3 years ago
Hi there @denijane,
Does including the line
from mpi4py import MPI
at the start of the script fix the problem?
Another thing that's not working is >output which doesn't print out in the notebook. But it's visible in konsole and in the text files, so that doesn't bother me that much. But the kernel crash is a problem because it's preventing me from combining two plots.
The output is legacy code. I would recommend looking into anesthetic for post-processing nested sampling tools
Hi there @denijane,
Does including the line
from mpi4py import MPI
at the start of the script fix the problem?
Hi William, thank you for the advise. It solved my problem. Now everything is working.
I also figured how to plot from chains and to save my posteriors and plot from them too, so everything is working fine.
This is just to note that I needed to modify the example run_pypolychord.py
to have the import at the top (and I needed to install the mpi4py module).
When I try to run two samplings in a row (for example with different priors), I get a kernel crash and I see in the console the following error:
I'm using Python 3.6 on Sabayon Linux. Another thing that's not working is >output which doesn't print out in the notebook. But it's visible in konsole and in the text files, so that doesn't bother me that much. But the kernel crash is a problem because it's preventing me from combining two plots.
If anyone could advise me about how to read from the saved files and produces the corner plots, that's a workaround. I can run 1 sampling, restart the kernel and run the next one. This is what i've done so far anyway. I saw that pymc can load the saved data and work with it, but I couldn't find this in polychord documentation.
P.S. I'm a newbie with polychord so I applogise if there is anything stupid. But the MPI problem is real.