# 'int' object has no attribute 'type' (pymc3)

**URL:** <https://discourse.pymc.io/t/int-object-has-no-attribute-type-pymc3/5484>\
**Category:** Questions\
**Created:** [July 20, 2020, 8:26pm UTC](https://discourse.pymc.io/t/int-object-has-no-attribute-type-pymc3/5484 "2020-07-20T20:26:37Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![sdeo](https://avatars.discourse-cdn.com/v4/letter/s/8491ac/32.png) [@sdeo](https://discourse.pymc.io/u/sdeo)\
**Post date:** [July 20, 2020, 8:26pm UTC](https://discourse.pymc.io/t/int-object-has-no-attribute-type-pymc3/5484/1 "2020-07-20T20:26:37Z")

</div>

Dear all,

The following code gives an error ‘int’ object has no attribute ‘type’ (pymc3) , because I am new in Theano based implementation in pymc3, I am not sure about the mistake. Maybe you could point it out fast. I have used the same code modified for pymc2, which worked well. Following is a snippet of the code, where some defined functions are outside this code.

# %%

import pymc3 as pm  
import theano.tensor as t  
import theano

exp\_rxn\_order\_CO = 0.840295  
exp\_rxn\_order\_O2 = 0.02  
exp\_act\_barrier = 0.69 # in eV

# define priors

# mu --\> Mean of the normal distribution

# tau = 1 / (St dev)^2

@theano.compile.ops.as\_op(itypes=[t.dscalar, t.dscalar, t.bscalar, t.dvector],otypes=[t.fscalar])  
def rxnOrderCO(dE\_CO\_ads,Ea\_CO\_oxid,T,P):  
rxnOrderCO = rxn\_order\_CO\_corrected(T,P,dE\_CO\_ads,Ea\_CO\_oxid)  
return rxnOrderCO

@theano.compile.ops.as\_op(itypes=[t.dscalar, t.dscalar, t.bscalar, t.dvector],otypes=[t.fscalar])  
def rxnOrderO2(dE\_CO\_ads,Ea\_CO\_oxid,T,P):  
rxnOrderO2 = rxn\_order\_O2\_corrected(T,P,dE\_CO\_ads,Ea\_CO\_oxid)  
return rxnOrderO2

@theano.compile.ops.as\_op(itypes=[t.dscalar, t.dscalar, t.bscalar, t.dvector],otypes=[t.fscalar])  
def act\_barrier(dE\_CO\_ads,Ea\_CO\_oxid,T,P):  
act\_barrier = apparent\_barrier\_corrected(T,P,dE\_CO\_ads,Ea\_CO\_oxid)  
return act\_barrier

with pm.Model() as model:  
T = 473  
P = [0.02, 0.2]  
dE\_CO\_ads = pm.Normal(‘dE\_CO\_ads’, mu=0, tau=1/0.1**2)  
Ea\_CO\_oxid = pm.Normal(‘Ea\_CO\_oxid’, mu=0.60, tau=1/0.1**2)

with pm.Model() as model:

```
rxnOrderCO = pm.Deterministic("rxnOrderCO", 
rxnOrderCO(dE_CO_ads,Ea_CO_oxid,T,P))    

#define likelihood
rxnOrder_CO = pm.Normal('rxnOrder_CO',mu=rxnOrderCO, tau = 1/0.1**2, value = exp_rxn_order_CO, observed=True)
rxnOrder_O2 = pm.Normal('rxnOrder_O2',mu=rxnOrderO2, tau = 1/0.1**2, value = exp_rxn_order_O2, observed=True)
actBarrier = pm.Normal('actBarrier',mu=act_barrier, tau = 1/0.1**2, value = exp_act_barrier, observed=True)
#%% Bayesian inference

```

Following is the error trace:

AttributeError Traceback (most recent call last)  
 in   
1 with pm.Model() as model:  
2  
----\> 3 rxnOrderCO = pm.Deterministic(“rxnOrderCO”, rxnOrderCO(dE\_CO\_ads,Ea\_CO\_oxid,T,P))  
4  
5 #define likelihood

C:\ProgramData\Anaconda3\lib\site-packages\theano\gof\op.py in **call** (self, \*inputs, \*\*kwargs)  
613 “”"  
614 return\_list = kwargs.pop(‘return\_list’, False)  
–\> 615 node = self.make\_node(\*inputs, \*\*kwargs)  
616  
617 if config.compute\_test\_value != ‘off’:

C:\ProgramData\Anaconda3\lib\site-packages\theano\gof\op.py in make\_node(self, \*inputs)  
981 raise ValueError(“We expected %d inputs but got %d.” %  
982 (len(self.itypes), len(inputs)))  
–\> 983 if not all(inp.type == it for inp, it in zip(inputs, self.itypes)):  
984 raise TypeError(  
985 "We expected inputs of types ‘%s’ but got types ‘%s’ " %

C:\ProgramData\Anaconda3\lib\site-packages\theano\gof\op.py in (.0)  
981 raise ValueError(“We expected %d inputs but got %d.” %  
982 (len(self.itypes), len(inputs)))  
–\> 983 if not all(inp.type == it for inp, it in zip(inputs, self.itypes)):  
984 raise TypeError(  
985 "We expected inputs of types ‘%s’ but got types ‘%s’ " %

AttributeError: ‘int’ object has no attribute ‘type’

I guess this has to do with the theono variables defined decorator for the functions. Thanks in advance.

---

<div class="post-metadata">

**Author:** ![ckrapu](https://yyz2.discourse-cdn.com/flex036/user_avatar/discourse.pymc.io/ckrapu/32/1366_2.png) [@ckrapu](https://discourse.pymc.io/u/ckrapu)\
**Post date:** [July 20, 2020, 10:44pm UTC](https://discourse.pymc.io/t/int-object-has-no-attribute-type-pymc3/5484/2 "2020-07-20T22:44:03Z")

</div>

What happens when you apply `theano.tensor.as_tensor_variable` to the second argument of your deterministic node?