Hi! I am trying to use estimate the parameters of a surface given reflectance data using a physical scattering model (mostly a complicated function). Everything works fine until the sampling, in which this error raises, TypeError: Elemwise{real,no_inplace}.grad illegally returned an integer-valued…

Sadly, I wasn’t able to make this kind of problem work in PyMC3. To be fair, it is not a problem of PyMC3, since it seems that is Theano the one that cannot handle complex numbers, even if both input and output of the forward function are real numbers (like in my case also). I was unwilling to write…

Using PyMC3 in a model with complex variables inside

Questions

leotercas April 11, 2023, 1:17pm 9

Hello Jean-François,

I would like to know if you found a solution to build a model with complex variables and also use NUTS from pymc3. I’m having a problem similar to what you once had.

Best regards,
Leo.

Incorperating numerical approximations into model

Topic		Replies	Views
Elemwise{abs,no_inplace}.grad error for Multinomial likelihood v5	3	261	October 26, 2023
Difficulty calculating model gradients, shared variables with static values, and sparse matrices Questions	4	952	March 18, 2019
Error in basic model due to trigonometric functions using math library Questions	10	2585	March 8, 2019
Pm.math or theano for trigonometric functions Questions	1	719	April 5, 2019
TypeError: Unknown parameter type: <class 'theano.tensor.var.TensorVariable'> Questions	2	3190	February 11, 2021

Using PyMC3 in a model with complex variables inside

Related topics