Kernels not copied when model cloned

[marpulli]

Describe the bug
When a model is cloned, any kernel attached to the graph is not copied. This means that cloned graphs with GP modules don't work.

To Reproduce
This is a unit test I added to gpregression_test.py file, it currently fails:

    def test_module_clone(self):
        D, X, Y, noise_var, lengthscale, variance = self.gen_data()
        dtype = 'float64'

        # Predict from original model
        m = self.gen_mxfusion_model(dtype, D, noise_var, lengthscale, variance)

        observed = [m.X, m.Y]
        infr = Inference(MAP(model=m, observed=observed), dtype=dtype)

        loss, _ = infr.run(X=mx.nd.array(X, dtype=dtype), Y=mx.nd.array(Y, dtype=dtype), max_iter=1)

        infr2 = TransferInference(ModulePredictionAlgorithm(m, observed=[m.X], target_variables=[m.Y]),
                                  infr_params=infr.params, dtype=np.float64)
        infr2.inference_algorithm.model.Y.factor.gp_predict.diagonal_variance = False
        infr2.inference_algorithm.model.Y.factor.gp_predict.noise_free = False
        res = infr2.run(X=mx.nd.array(X, dtype=dtype))[0]
        mu_mf, var_mf = res[0].asnumpy()[0], res[1].asnumpy()[0]

        # Clone model
        cloned_model = m.clone()

        # Predict from cloned model
        observed = [cloned_model.X, cloned_model.Y]
        infr = Inference(MAP(model=cloned_model, observed=observed), dtype=dtype)

        loss, _ = infr.run(X=mx.nd.array(X, dtype=dtype), Y=mx.nd.array(Y, dtype=dtype), max_iter=1)

        infr2_clone = TransferInference(ModulePredictionAlgorithm(cloned_model, observed=[cloned_model.X],
                                                                  target_variables=[cloned_model.Y]),
                                        infr_params=infr.params, dtype=np.float64)
        infr2_clone.inference_algorithm.model.Y.factor.gp_predict.diagonal_variance = False
        infr2_clone.inference_algorithm.model.Y.factor.gp_predict.noise_free = False
        res = infr2_clone.run(X=mx.nd.array(X, dtype=dtype))[0]
        mu_mf_clone, var_mf_clone = res[0].asnumpy()[0], res[1].asnumpy()[0]

        assert np.allclose(mu_mf, mu_mf_clone)
        assert np.allclose(var_mf, var_mf_clone)

Error message:

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 
../../mxfusion/inference/inference.py:171: in run
    return executor(mx.nd.zeros(1, ctx=self.mxnet_context), *data)
../../../../miniconda3/lib/python3.6/site-packages/mxnet/gluon/block.py:540: in __call__
    out = self.forward(*args)
../../../../miniconda3/lib/python3.6/site-packages/mxnet/gluon/block.py:917: in forward
    return self.hybrid_forward(ndarray, x, *args, **params)
../../mxfusion/inference/inference_alg.py:83: in hybrid_forward
    obj = self._infr_method.compute(F=F, variables=variables)
../../mxfusion/inference/map.py:83: in compute
    logL = self.model.log_pdf(F=F, variables=variables)
../../mxfusion/models/factor_graph.py:234: in log_pdf
    F=F, variables=variables, targets=module_targets)))
../../mxfusion/modules/module.py:321: in log_pdf
    result = alg.compute(F, variables)
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ 

self = <mxfusion.modules.gp_modules.gp_regression.GPRegressionLogPdf object at 0x122e384e0>
F = <module 'mxnet.ndarray' from '/Users/marpulli/miniconda3/lib/python3.6/site-packages/mxnet/ndarray/__init__.py'>
variables = {'271f2355_72ae_48a2_b46a_ff58815c1f07': 
[[[0.26455561 0.77423369]
  [0.45615033 0.56843395]
  [0.0187898  0.6176355 ....06185717 -0.01682754
    0.03680881  0.0380233  -0.02780953  0.03818348 -0.04858983]]]
<NDArray 1x10x10 @cpu(0)>, ...}

    def compute(self, F, variables):
        has_mean = self.model.F.factor.has_mean
        X = variables[self.model.X]
        Y = variables[self.model.Y]
        noise_var = variables[self.model.noise_var]
        D = Y.shape[-1]
        N = X.shape[-2]
>       kern = self.model.kernel
E       AttributeError: 'Model' object has no attribute 'kernel'

../../mxfusion/modules/gp_modules/gp_regression.py:49: AttributeError

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

Desktop (please complete the following information):

• OS: iOS
• Python version3.6
• MXNet version 1.3
• MXFusion version master
• MXNet contextcpu
• MXNet dtype float64

Additional context
Add any other context about the problem here.