In my script, I perform a fitting with lmfit
the fitting routine called 80 times the solver to compare calculation with experimental (observed data) this took 20 seconds.
After the fitting, I want to execute PyMC3
using theano
to determine the posterior distributions for each model parameter. Here I am using pm.Slice()
and each call is demanding 10 seconds…
I want to know if this has already happened before, and how to speed up?
Thanks.-
Ps.: My function has a try, except statement. The “failed” case returns some vale out of sense, to avoid digging in that samples