Error when assign model variable as kernel parameters

[zhenwendai]

Describe the bug
Error when assign model variable as kernel parameters.

from mxfusion import Model, Variable
from mxfusion.components.variables import PositiveTransformation, Softplus
from mxfusion.components.distributions.gp.kernels import RBF
from mxfusion.modules.gp_modules import SparseGPRegression, GPRegression

m = Model()
m.N = Variable()
m.X = Variable(shape=(m.N, X_prepared.shape[-1]))
m.noise_var = Variable(shape=(1,), transformation=PositiveTransformation(), initial_value=0.01)
m.lengthscale = Variable(shape=(X_prepared.shape[-1],), transformation=Softplus(0.01), 
                         initial_value=mx.nd.array([1]*X_prepared.shape[-1], dtype='float64'))
m.kernel = RBF(input_dim=X_prepared.shape[-1], variance=1, lengthscale=m.lengthscale, ARD=True)
m.Y = GPRegression.define_variable(X=m.X, kernel=m.kernel, noise_var=m.noise_var, shape=(m.N, Y_prepared.shape[-1]))

To Reproduce
Steps to reproduce the behavior:

1. Build a model like '...'
2. Run inference '...'
3. See error

Preferably this is in the form of a small snippet of runnable code that reproduces the behavior.

Expected behavior
A clear and concise description of what you expected to happen. A

Desktop (please complete the following information):

• OS: MacOS
• Python version 3.6
• MXNet version 1.3
• MXFusion version 0.3.0
• MXNet context CPU
• MXNet dtype float64

Additional context
Add any other context about the problem here.