I came late to the recent “latest” version of PyMC3, I was working before with PyMC3 (version 3.8), when the next statements were still used:
import theano.tensor as tt
from theano.compile.ops import as_op
Now I have aesara 2.0.7
, pymc3 3.11.2
, and Theano-PyMC 1.1.2
.
I am using PyMC3 to do Bayesian Inversion of a model with data… do I still need to do stuff like follows:
@as_op(itypes=[tt.dscalar,tt.dscalar,tt.dscalar,tt.dscalar,tt.dscalar], otypes=[tt.dvector])
def th_forward_model(param1,param2,param3,param4,param5):
return model.vec_func(param1,param2,param3,param4,param5)[1]
Any advice will be gratefully acknowledged.
Even here the authors used the old syntax, Link to tutorial
In addition to this, I get the following error:
import pymc3 as pm
File "/home/valdez/anaconda3/lib/python3.7/site-packages/pymc3/__init__.py", line 23, in <module>
import theano
ModuleNotFoundError: No module named 'theano'
Thanks in advance and my apologies if this has been already asked.