search

joshjchayes / TransitFit

Transit light curve fitting using nested sampling
GNU General Public License v3.0
4 stars 1 forks source link

LinAlgError for large Single run #2

Closed JStMorgan closed 4 years ago

JStMorgan commented 4 years ago

Hello Josh,

I've been using TransitFit for a combined WASP-43b data set (26 of our epochs and 16 optical spectrographic), using the Single LD fitting mode for a total of 152 parameters to be retrieved. At some point in the run, it failed with the following error message:

LinAlgError: 152-th leading minor of the array is not positive definite

From my brief reading on this, it seems to be an error in the covariance matrix, with a value becoming negative which shouldn't be. I still have the full error output in Spyder - will you be around on Monday to check it out?

Many thanks, Jake

joshjchayes commented 4 years ago

This issue was due to a bug in dynesty. To my knowledge this is now fixed.