If scan is too difficult to use, I can do that. I’m running into a bit of a problem with the as_tensor_variable function when I do this. Ideally, I would like to stuff my phi_pymc variables into a list, then turn that list into a tensor. So, something like:
p0 = [tnd, v, sigma, a[0]]
p1 = [tnd, v, sigma, a[1]]
p2 = [tnd, v, sigma, a[2]]
phi_pymc = at.as_tensor_variable([p0, p1, p2])
This doesn’t work however because I can’t convert objects into tensor variables. I get the following error:
TypeError: Unsupported dtype for TensorType: object
Is there a way to get around this, or do I just have to shove all priors into a single long list?