Reputation: 103
i hope you are having a better day than i do. I am developing a project that is due within a week and the codebase that i am working on is no longer working. It was working completely fine a month ago with 'Pymc3' but now 'Pymc3' is not working due to problems with Theano. So i have decided to switch to Pymc but i am getting and error as stated in the title.
Here is my code :
import pymc as pm
with pm.Model() as gaussian:
a = pm.Normal('a', mu=0, sd=10) #sigma
b = pm.Normal('b', mu=0, sd=1) #intercept #m
e = pm.HalfCauchy('e', 5) #slope
u = pm.Deterministic('u', a + b * arr) #acikla #mx+b
yPred = pm.Normal('yPred', mu=u, sd=e, observed=arr2) #likelihood
trace_g = pm.sample(2000, tune=1000) #posterior
I honestly don't know what is wrong. I have checked the documentation but i couldn't manage to find anything so any help would be much much appreciated.
Upvotes: 0
Views: 187
Reputation: 103
Okay apparently they have decided to change previous parameter 'sd' to 'sigma' now. I have replaced everywhere that it says 'sd' to 'sigma' and problem has resolved.
Upvotes: 0