diff --git a/.travis.yml b/.travis.yml
index 9a738047..7897b009 100644
--- a/.travis.yml
+++ b/.travis.yml
@@ -29,6 +29,7 @@ install:
 - echo $PATH
 - source install_retry.sh
 - pip install codecov
+- pip install coveralls
 - pip install pypandoc
 - pip install git+git://github.com/BRML/climin.git
 - pip install autograd
@@ -40,6 +41,7 @@ script:
 
 after_success:
   - codecov
+  - coveralls
 
 before_deploy:
   - cd doc
diff --git a/GPy/kern/src/kern.py b/GPy/kern/src/kern.py
index 4379fb71..c6f62bd9 100644
--- a/GPy/kern/src/kern.py
+++ b/GPy/kern/src/kern.py
@@ -211,6 +211,12 @@ class Kern(Parameterized):
     def input_sensitivity(self, summarize=True):
         """
         Returns the sensitivity for each dimension of this kernel.
+
+        This is an arbitrary measurement based on the parameters
+        of the kernel per dimension and scaling in general.
+
+        Use this as relative measurement, not for absolute comparison between
+        kernels.
         """
         return np.zeros(self.input_dim)
 
diff --git a/GPy/kern/src/prod.py b/GPy/kern/src/prod.py
index 674365ba..ab3df431 100644
--- a/GPy/kern/src/prod.py
+++ b/GPy/kern/src/prod.py
@@ -99,7 +99,7 @@ class Prod(CombinationKernel):
 
     def input_sensitivity(self, summarize=True):
         if summarize:
-            i_s = np.zeros((self.input_dim))
+            i_s = np.ones((self.input_dim))
             for k in self.parts:
                 i_s[k._all_dims_active] *= k.input_sensitivity(summarize)
             return i_s
diff --git a/GPy/kern/src/stationary.py b/GPy/kern/src/stationary.py
index 1ce8084f..b6b13d47 100644
--- a/GPy/kern/src/stationary.py
+++ b/GPy/kern/src/stationary.py
@@ -51,6 +51,10 @@ class Stationary(Kern):
 
     The lengthscale(s) and variance parameters are added to the structure automatically.
 
+    Thanks to @strongh:
+    In Stationary, a covariance function is defined in GPy as stationary when it depends only on the l2-norm |x_1 - x_2 |. 
+    However this is the typical definition of isotropy, while stationarity is usually a bit more relaxed. 
+    The more common version of stationarity is that the covariance is a function of x_1 - x_2 (See e.g. R&W first paragraph of section 4.1).
     """
 
     def __init__(self, input_dim, variance, lengthscale, ARD, active_dims, name, useGPU=False):
diff --git a/GPy/mappings/additive.py b/GPy/mappings/additive.py
index 1c86b680..0fb72ca4 100644
--- a/GPy/mappings/additive.py
+++ b/GPy/mappings/additive.py
@@ -23,9 +23,10 @@ class Additive(Mapping):
         assert(mapping1.input_dim==mapping2.input_dim)
         assert(mapping1.output_dim==mapping2.output_dim)
         input_dim, output_dim = mapping1.input_dim, mapping1.output_dim
-        Mapping.__init__(self, input_dim=input_dim, output_dim=output_dim)
+        super(Additive, self).__init__(input_dim=input_dim, output_dim=output_dim)
         self.mapping1 = mapping1
         self.mapping2 = mapping2
+        self.link_parameters(self.mapping1, self.mapping2)
 
     def f(self, X):
         return self.mapping1.f(X) + self.mapping2.f(X)
diff --git a/GPy/mappings/linear.py b/GPy/mappings/linear.py
index c105dd81..26504830 100644
--- a/GPy/mappings/linear.py
+++ b/GPy/mappings/linear.py
@@ -33,7 +33,7 @@ class Linear(Mapping):
         return np.dot(X, self.A)
 
     def update_gradients(self, dL_dF, X):
-        self.A.gradient = np.dot( X.T, dL_dF)
+        self.A.gradient = np.dot(X.T, dL_dF)
 
     def gradients_X(self, dL_dF, X):
         return np.dot(dL_dF, self.A.T)
diff --git a/GPy/testing/b-gplvm-save.npz b/GPy/testing/b-gplvm-save.npz
new file mode 100644
index 00000000..bcbe84d7
Binary files /dev/null and b/GPy/testing/b-gplvm-save.npz differ
diff --git a/GPy/testing/baseline/bayesian_gplvm_gradient.npz b/GPy/testing/baseline/bayesian_gplvm_gradient.npz
index f3b56c49..b48db887 100644
Binary files a/GPy/testing/baseline/bayesian_gplvm_gradient.npz and b/GPy/testing/baseline/bayesian_gplvm_gradient.npz differ
diff --git a/GPy/testing/baseline/bayesian_gplvm_inducing.npz b/GPy/testing/baseline/bayesian_gplvm_inducing.npz
index 2ad1866a..82e44f54 100644
Binary files a/GPy/testing/baseline/bayesian_gplvm_inducing.npz and b/GPy/testing/baseline/bayesian_gplvm_inducing.npz differ
diff --git a/GPy/testing/baseline/bayesian_gplvm_inducing_3d.npz b/GPy/testing/baseline/bayesian_gplvm_inducing_3d.npz
index 02c89096..e55b4b01 100644
Binary files a/GPy/testing/baseline/bayesian_gplvm_inducing_3d.npz and b/GPy/testing/baseline/bayesian_gplvm_inducing_3d.npz differ
diff --git a/GPy/testing/baseline/bayesian_gplvm_latent.npz b/GPy/testing/baseline/bayesian_gplvm_latent.npz
index 3fba9c06..3e27bc64 100644
Binary files a/GPy/testing/baseline/bayesian_gplvm_latent.npz and b/GPy/testing/baseline/bayesian_gplvm_latent.npz differ
diff --git a/GPy/testing/baseline/bayesian_gplvm_latent_3d.npz b/GPy/testing/baseline/bayesian_gplvm_latent_3d.npz
index a4f4edad..b4da4571 100644
Binary files a/GPy/testing/baseline/bayesian_gplvm_latent_3d.npz and b/GPy/testing/baseline/bayesian_gplvm_latent_3d.npz differ
diff --git a/GPy/testing/baseline/bayesian_gplvm_magnification.npz b/GPy/testing/baseline/bayesian_gplvm_magnification.npz
index cd5ff303..dd83f22d 100644
Binary files a/GPy/testing/baseline/bayesian_gplvm_magnification.npz and b/GPy/testing/baseline/bayesian_gplvm_magnification.npz differ
diff --git a/GPy/testing/baseline/coverage_3d_plot.npz b/GPy/testing/baseline/coverage_3d_plot.npz
index cdf55bc8..6372cb73 100644
Binary files a/GPy/testing/baseline/coverage_3d_plot.npz and b/GPy/testing/baseline/coverage_3d_plot.npz differ
diff --git a/GPy/testing/baseline/coverage_annotation_interact.npz b/GPy/testing/baseline/coverage_annotation_interact.npz
index 437ad498..16fd4e38 100644
Binary files a/GPy/testing/baseline/coverage_annotation_interact.npz and b/GPy/testing/baseline/coverage_annotation_interact.npz differ
diff --git a/GPy/testing/baseline/coverage_gradient.npz b/GPy/testing/baseline/coverage_gradient.npz
index 32eaf6c6..a55dd97b 100644
Binary files a/GPy/testing/baseline/coverage_gradient.npz and b/GPy/testing/baseline/coverage_gradient.npz differ
diff --git a/GPy/testing/baseline/coverage_imshow_interact.npz b/GPy/testing/baseline/coverage_imshow_interact.npz
index 78dd7755..55dd013c 100644
Binary files a/GPy/testing/baseline/coverage_imshow_interact.npz and b/GPy/testing/baseline/coverage_imshow_interact.npz differ
diff --git a/GPy/testing/baseline/gp_2d_data.npz b/GPy/testing/baseline/gp_2d_data.npz
index 773660eb..0c703f69 100644
Binary files a/GPy/testing/baseline/gp_2d_data.npz and b/GPy/testing/baseline/gp_2d_data.npz differ
diff --git a/GPy/testing/baseline/gp_2d_in_error.npz b/GPy/testing/baseline/gp_2d_in_error.npz
index 00ca4eca..beb00720 100644
Binary files a/GPy/testing/baseline/gp_2d_in_error.npz and b/GPy/testing/baseline/gp_2d_in_error.npz differ
diff --git a/GPy/testing/baseline/gp_2d_inducing.npz b/GPy/testing/baseline/gp_2d_inducing.npz
index ce6404fc..2a249e27 100644
Binary files a/GPy/testing/baseline/gp_2d_inducing.npz and b/GPy/testing/baseline/gp_2d_inducing.npz differ
diff --git a/GPy/testing/baseline/gp_2d_mean.npz b/GPy/testing/baseline/gp_2d_mean.npz
index 238dc028..dd9ef64d 100644
Binary files a/GPy/testing/baseline/gp_2d_mean.npz and b/GPy/testing/baseline/gp_2d_mean.npz differ
diff --git a/GPy/testing/baseline/gp_3d_data.npz b/GPy/testing/baseline/gp_3d_data.npz
index 1f52eab4..e190118a 100644
Binary files a/GPy/testing/baseline/gp_3d_data.npz and b/GPy/testing/baseline/gp_3d_data.npz differ
diff --git a/GPy/testing/baseline/gp_3d_inducing.npz b/GPy/testing/baseline/gp_3d_inducing.npz
index 676a16e5..b6316363 100644
Binary files a/GPy/testing/baseline/gp_3d_inducing.npz and b/GPy/testing/baseline/gp_3d_inducing.npz differ
diff --git a/GPy/testing/baseline/gp_3d_mean.npz b/GPy/testing/baseline/gp_3d_mean.npz
index 1a5eed42..ae0a36ca 100644
Binary files a/GPy/testing/baseline/gp_3d_mean.npz and b/GPy/testing/baseline/gp_3d_mean.npz differ
diff --git a/GPy/testing/baseline/gp_class_likelihood.npz b/GPy/testing/baseline/gp_class_likelihood.npz
index 6db3e8ff..288b15ef 100644
Binary files a/GPy/testing/baseline/gp_class_likelihood.npz and b/GPy/testing/baseline/gp_class_likelihood.npz differ
diff --git a/GPy/testing/baseline/gp_class_raw.npz b/GPy/testing/baseline/gp_class_raw.npz
index f49a775f..08bb56df 100644
Binary files a/GPy/testing/baseline/gp_class_raw.npz and b/GPy/testing/baseline/gp_class_raw.npz differ
diff --git a/GPy/testing/baseline/gp_class_raw_link.npz b/GPy/testing/baseline/gp_class_raw_link.npz
index 231adca7..01490f4c 100644
Binary files a/GPy/testing/baseline/gp_class_raw_link.npz and b/GPy/testing/baseline/gp_class_raw_link.npz differ
diff --git a/GPy/testing/baseline/gp_conf.npz b/GPy/testing/baseline/gp_conf.npz
index d178b5c0..3eaf087e 100644
Binary files a/GPy/testing/baseline/gp_conf.npz and b/GPy/testing/baseline/gp_conf.npz differ
diff --git a/GPy/testing/baseline/gp_data.npz b/GPy/testing/baseline/gp_data.npz
index fef7d0f5..2a4606df 100644
Binary files a/GPy/testing/baseline/gp_data.npz and b/GPy/testing/baseline/gp_data.npz differ
diff --git a/GPy/testing/baseline/gp_density.npz b/GPy/testing/baseline/gp_density.npz
index 9bf3d757..6a3235fa 100644
Binary files a/GPy/testing/baseline/gp_density.npz and b/GPy/testing/baseline/gp_density.npz differ
diff --git a/GPy/testing/baseline/gp_in_error.npz b/GPy/testing/baseline/gp_in_error.npz
index eb693591..072fbc12 100644
Binary files a/GPy/testing/baseline/gp_in_error.npz and b/GPy/testing/baseline/gp_in_error.npz differ
diff --git a/GPy/testing/baseline/gp_mean.npz b/GPy/testing/baseline/gp_mean.npz
index e47dad83..432b2c25 100644
Binary files a/GPy/testing/baseline/gp_mean.npz and b/GPy/testing/baseline/gp_mean.npz differ
diff --git a/GPy/testing/baseline/gp_out_error.npz b/GPy/testing/baseline/gp_out_error.npz
index 967e322d..4e86a06d 100644
Binary files a/GPy/testing/baseline/gp_out_error.npz and b/GPy/testing/baseline/gp_out_error.npz differ
diff --git a/GPy/testing/baseline/gp_samples.npz b/GPy/testing/baseline/gp_samples.npz
index 5f19e9c4..774a576b 100644
Binary files a/GPy/testing/baseline/gp_samples.npz and b/GPy/testing/baseline/gp_samples.npz differ
diff --git a/GPy/testing/baseline/gplvm_gradient.npz b/GPy/testing/baseline/gplvm_gradient.npz
index 14ca0cfd..fb6a1658 100644
Binary files a/GPy/testing/baseline/gplvm_gradient.npz and b/GPy/testing/baseline/gplvm_gradient.npz differ
diff --git a/GPy/testing/baseline/gplvm_latent.npz b/GPy/testing/baseline/gplvm_latent.npz
index fcd07f97..fac4cca9 100644
Binary files a/GPy/testing/baseline/gplvm_latent.npz and b/GPy/testing/baseline/gplvm_latent.npz differ
diff --git a/GPy/testing/baseline/gplvm_latent_3d.npz b/GPy/testing/baseline/gplvm_latent_3d.npz
index 448362ae..2fe42704 100644
Binary files a/GPy/testing/baseline/gplvm_latent_3d.npz and b/GPy/testing/baseline/gplvm_latent_3d.npz differ
diff --git a/GPy/testing/baseline/gplvm_magnification.npz b/GPy/testing/baseline/gplvm_magnification.npz
index 7275a82a..fb84f382 100644
Binary files a/GPy/testing/baseline/gplvm_magnification.npz and b/GPy/testing/baseline/gplvm_magnification.npz differ
diff --git a/GPy/testing/baseline/kern_ARD.npz b/GPy/testing/baseline/kern_ARD.npz
index 77bedd3d..7735251e 100644
Binary files a/GPy/testing/baseline/kern_ARD.npz and b/GPy/testing/baseline/kern_ARD.npz differ
diff --git a/GPy/testing/baseline/kern_cov_1d.npz b/GPy/testing/baseline/kern_cov_1d.npz
index 285e2e6d..3bb0ac83 100644
Binary files a/GPy/testing/baseline/kern_cov_1d.npz and b/GPy/testing/baseline/kern_cov_1d.npz differ
diff --git a/GPy/testing/baseline/kern_cov_2d.npz b/GPy/testing/baseline/kern_cov_2d.npz
index d65d1d65..78a5930b 100644
Binary files a/GPy/testing/baseline/kern_cov_2d.npz and b/GPy/testing/baseline/kern_cov_2d.npz differ
diff --git a/GPy/testing/baseline/kern_cov_3d.npz b/GPy/testing/baseline/kern_cov_3d.npz
index b98bda42..a76f302f 100644
Binary files a/GPy/testing/baseline/kern_cov_3d.npz and b/GPy/testing/baseline/kern_cov_3d.npz differ
diff --git a/GPy/testing/baseline/kern_cov_no_lim.npz b/GPy/testing/baseline/kern_cov_no_lim.npz
index aa4b9512..019db745 100644
Binary files a/GPy/testing/baseline/kern_cov_no_lim.npz and b/GPy/testing/baseline/kern_cov_no_lim.npz differ
diff --git a/GPy/testing/baseline/sparse_gp_class_likelihood.npz b/GPy/testing/baseline/sparse_gp_class_likelihood.npz
index 5aee4612..8d9f1731 100644
Binary files a/GPy/testing/baseline/sparse_gp_class_likelihood.npz and b/GPy/testing/baseline/sparse_gp_class_likelihood.npz differ
diff --git a/GPy/testing/baseline/sparse_gp_class_raw.npz b/GPy/testing/baseline/sparse_gp_class_raw.npz
index 7c9404c8..d6634501 100644
Binary files a/GPy/testing/baseline/sparse_gp_class_raw.npz and b/GPy/testing/baseline/sparse_gp_class_raw.npz differ
diff --git a/GPy/testing/baseline/sparse_gp_class_raw_link.npz b/GPy/testing/baseline/sparse_gp_class_raw_link.npz
index 9512e9e7..0d09f5ae 100644
Binary files a/GPy/testing/baseline/sparse_gp_class_raw_link.npz and b/GPy/testing/baseline/sparse_gp_class_raw_link.npz differ
diff --git a/GPy/testing/baseline/sparse_gp_data_error.npz b/GPy/testing/baseline/sparse_gp_data_error.npz
index beaea488..d6396b8a 100644
Binary files a/GPy/testing/baseline/sparse_gp_data_error.npz and b/GPy/testing/baseline/sparse_gp_data_error.npz differ
diff --git a/GPy/testing/meanfunc_tests.py b/GPy/testing/meanfunc_tests.py
index 815c024f..53482a7a 100644
--- a/GPy/testing/meanfunc_tests.py
+++ b/GPy/testing/meanfunc_tests.py
@@ -28,10 +28,49 @@ class MFtests(unittest.TestCase):
         A linear mean function with parameters that we'll learn alongside the kernel
         """
 
+        X = np.linspace(-1,10,50).reshape(-1,1)
+        
+        Y = 3-np.abs((X-6))
+        Y += .5*np.cos(3*X) + 0.3*np.random.randn(*X.shape) 
+
+        mf = GPy.mappings.PiecewiseLinear(1, 1, [-1,1], [9,2])
+
+        k =GPy.kern.RBF(1)
+        lik = GPy.likelihoods.Gaussian()
+        m = GPy.core.GP(X, Y, kernel=k, likelihood=lik, mean_function=mf)
+        self.assertTrue(m.checkgrad())
+
+    def test_parametric_mean_function_composition(self):
+        """
+        A linear mean function with parameters that we'll learn alongside the kernel
+        """
+
         X = np.linspace(0,10,50).reshape(-1,1)
         Y = np.sin(X) + 0.5*np.cos(3*X) + 0.1*np.random.randn(*X.shape) + 3*X
 
-        mf = GPy.mappings.Linear(1,1)
+        mf = GPy.mappings.Compound(GPy.mappings.Linear(1,1), 
+                                   GPy.mappings.Kernel(1, 1, np.random.normal(0,1,(1,1)), 
+                                                       GPy.kern.RBF(1))
+                                   )
+
+        k =GPy.kern.RBF(1)
+        lik = GPy.likelihoods.Gaussian()
+        m = GPy.core.GP(X, Y, kernel=k, likelihood=lik, mean_function=mf)
+        self.assertTrue(m.checkgrad())
+
+    def test_parametric_mean_function_additive(self):
+        """
+        A linear mean function with parameters that we'll learn alongside the kernel
+        """
+
+        X = np.linspace(0,10,50).reshape(-1,1)
+        Y = np.sin(X) + 0.5*np.cos(3*X) + 0.1*np.random.randn(*X.shape) + 3*X
+
+        mf = GPy.mappings.Additive(GPy.mappings.Constant(1,1,3),
+               GPy.mappings.Additive(GPy.mappings.MLP(1,1),
+                     GPy.mappings.Identity(1,1)
+                           )
+                        )
 
         k =GPy.kern.RBF(1)
         lik = GPy.likelihoods.Gaussian()
diff --git a/GPy/testing/plotting_tests.py b/GPy/testing/plotting_tests.py
index 4922a3ec..f115c801 100644
--- a/GPy/testing/plotting_tests.py
+++ b/GPy/testing/plotting_tests.py
@@ -74,12 +74,13 @@ except ImportError:
 
 extensions = ['npz']
 
+basedir = os.path.dirname(os.path.relpath(os.path.abspath(__file__)))
+
 def _image_directories():
     """
     Compute the baseline and result image directories for testing *func*.
     Create the result directory if it doesn't exist.
     """
-    basedir = os.path.dirname(os.path.relpath(os.path.abspath(__file__)))
     #module_name = __init__.__module__
     #mods = module_name.split('.')
     #basedir = os.path.join(*mods)
@@ -349,7 +350,9 @@ def test_sparse():
     m = GPy.models.SparseGPRegression(X, Y, X_variance=np.ones_like(X)*0.1)
     #m.optimize()
     #m.plot_inducing()
-    m.plot_data()
+    _, ax = plt.subplots()
+    m.plot_data(ax=ax)
+    m.plot_data_error(ax=ax)
     for do_test in _image_comparison(baseline_images=['sparse_gp_{}'.format(sub) for sub in ['data_error']], extensions=extensions):
         yield (do_test, )
 
@@ -397,31 +400,39 @@ def test_sparse_classification():
         yield (do_test, )
 
 def test_gplvm():
-    from ..examples.dimensionality_reduction import _simulate_matern
-    from ..kern import RBF
-    from ..models import GPLVM
+    from GPy.models import GPLVM
     np.random.seed(12345)
     matplotlib.rcParams.update(matplotlib.rcParamsDefault)
     #matplotlib.rcParams[u'figure.figsize'] = (4,3)
     matplotlib.rcParams[u'text.usetex'] = False
-    Q = 3
+    #Q = 3
     # Define dataset
-    N = 10
-    k1 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[10,10,10,0.1,0.1]), ARD=True)
-    k2 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[10,0.1,10,0.1,10]), ARD=True)
-    k3 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[0.1,0.1,10,10,10]), ARD=True)
-    X = np.random.normal(0, 1, (N, 5))
-    A = np.random.multivariate_normal(np.zeros(N), k1.K(X), Q).T
-    B = np.random.multivariate_normal(np.zeros(N), k2.K(X), Q).T
-    C = np.random.multivariate_normal(np.zeros(N), k3.K(X), Q).T
+    #N = 60
+    #k1 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[10,10,10,0.1,0.1]), ARD=True)
+    #k2 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[10,0.1,10,0.1,10]), ARD=True)
+    #k3 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[0.1,0.1,10,10,10]), ARD=True)
+    #X = np.random.normal(0, 1, (N, 5))
+    #A = np.random.multivariate_normal(np.zeros(N), k1.K(X), Q).T
+    #B = np.random.multivariate_normal(np.zeros(N), k2.K(X), Q).T
+    #C = np.random.multivariate_normal(np.zeros(N), k3.K(X), Q).T
+    #Y = np.vstack((A,B,C))
+    #labels = np.hstack((np.zeros(A.shape[0]), np.ones(B.shape[0]), np.ones(C.shape[0])*2))
 
-    Y = np.vstack((A,B,C))
-    labels = np.hstack((np.zeros(A.shape[0]), np.ones(B.shape[0]), np.ones(C.shape[0])*2))
+    #k = RBF(Q, ARD=True, lengthscale=2)  # + kern.white(Q, _np.exp(-2)) # + kern.bias(Q)
+    pars = np.load(os.path.join(basedir, 'b-gplvm-save.npz'))
+    Y = pars['Y']
+    Q = pars['Q']
+    labels = pars['labels']
+
+    import warnings
+    with warnings.catch_warnings(record=True) as w:
+        warnings.simplefilter('always')  # always print
+        m = GPLVM(Y, Q, initialize=False)
+    m.update_model(False)
+    m.initialize_parameter()
+    m[:] = pars['gplvm_p']
+    m.update_model(True)
 
-    k = RBF(Q, ARD=True, lengthscale=2)  # + kern.white(Q, _np.exp(-2)) # + kern.bias(Q)
-    m = GPLVM(Y, Q, init="PCA", kernel=k)
-    m.kern.lengthscale[:] = [1./.3, 1./.1, 1./.7]
-    m.likelihood.variance = .001
     #m.optimize(messages=0)
     np.random.seed(111)
     m.plot_latent(labels=labels)
@@ -436,31 +447,40 @@ def test_gplvm():
         yield (do_test, )
 
 def test_bayesian_gplvm():
-    from ..examples.dimensionality_reduction import _simulate_matern
-    from ..kern import RBF
     from ..models import BayesianGPLVM
     np.random.seed(12345)
     matplotlib.rcParams.update(matplotlib.rcParamsDefault)
     #matplotlib.rcParams[u'figure.figsize'] = (4,3)
     matplotlib.rcParams[u'text.usetex'] = False
-    Q = 3
+    #Q = 3
     # Define dataset
-    N = 10
-    k1 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[10,10,10,0.1,0.1]), ARD=True)
-    k2 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[10,0.1,10,0.1,10]), ARD=True)
-    k3 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[0.1,0.1,10,10,10]), ARD=True)
-    X = np.random.normal(0, 1, (N, 5))
-    A = np.random.multivariate_normal(np.zeros(N), k1.K(X), Q).T
-    B = np.random.multivariate_normal(np.zeros(N), k2.K(X), Q).T
-    C = np.random.multivariate_normal(np.zeros(N), k3.K(X), Q).T
+    #N = 10
+    #k1 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[10,10,10,0.1,0.1]), ARD=True)
+    #k2 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[10,0.1,10,0.1,10]), ARD=True)
+    #k3 = GPy.kern.RBF(5, variance=1, lengthscale=1./np.random.dirichlet(np.r_[0.1,0.1,10,10,10]), ARD=True)
+    #X = np.random.normal(0, 1, (N, 5))
+    #A = np.random.multivariate_normal(np.zeros(N), k1.K(X), Q).T
+    #B = np.random.multivariate_normal(np.zeros(N), k2.K(X), Q).T
+    #C = np.random.multivariate_normal(np.zeros(N), k3.K(X), Q).T
 
-    Y = np.vstack((A,B,C))
-    labels = np.hstack((np.zeros(A.shape[0]), np.ones(B.shape[0]), np.ones(C.shape[0])*2))
+    #Y = np.vstack((A,B,C))
+    #labels = np.hstack((np.zeros(A.shape[0]), np.ones(B.shape[0]), np.ones(C.shape[0])*2))
+
+    #k = RBF(Q, ARD=True, lengthscale=2)  # + kern.white(Q, _np.exp(-2)) # + kern.bias(Q)
+    pars = np.load(os.path.join(basedir, 'b-gplvm-save.npz'))
+    Y = pars['Y']
+    Q = pars['Q']
+    labels = pars['labels']
+
+    import warnings
+    with warnings.catch_warnings(record=True) as w:
+        warnings.simplefilter('always')  # always print
+        m = BayesianGPLVM(Y, Q, initialize=False)
+    m.update_model(False)
+    m.initialize_parameter()
+    m[:] = pars['bgplvm_p']
+    m.update_model(True)
 
-    k = RBF(Q, ARD=True, lengthscale=2)  # + kern.white(Q, _np.exp(-2)) # + kern.bias(Q)
-    m = BayesianGPLVM(Y, Q, init="PCA", kernel=k)
-    m.kern.lengthscale[:] = [1./.3, 1./.1, 1./.7]
-    m.likelihood.variance = .001
     #m.optimize(messages=0)
     np.random.seed(111)
     m.plot_inducing(projection='2d')
diff --git a/GPy/util/datasets.py b/GPy/util/datasets.py
index 2d1d3244..af8912ac 100644
--- a/GPy/util/datasets.py
+++ b/GPy/util/datasets.py
@@ -98,7 +98,7 @@ def data_available(dataset_name=None):
     try:
         from itertools import zip_longest
     except ImportError:
-        from itertools import zip_longest as zip_longest
+        from itertools import izip_longest as zip_longest
     dr = data_resources[dataset_name]
     zip_urls = (dr['files'], )
     if 'save_names' in dr: zip_urls += (dr['save_names'], )
@@ -1033,14 +1033,18 @@ def singlecell_rna_seq_deng(dataset='singlecell_deng'):
                 data = inner.RPKM.to_frame()
                 data.columns = [file_info.name[:-18]]
                 gene_info = inner.Refseq_IDs.to_frame()
-                gene_info.columns = [file_info.name[:-18]]
+                gene_info.columns = ['NCBI Reference Sequence']
             else:
                 data[file_info.name[:-18]] = inner.RPKM
-                gene_info[file_info.name[:-18]] = inner.Refseq_IDs
+                #gene_info[file_info.name[:-18]] = inner.Refseq_IDs
 
     # Strip GSM number off data index
     rep = re.compile('GSM\d+_')
-    data.columns = data.columns.to_series().apply(lambda row: row[rep.match(row).end():])
+
+    from pandas import MultiIndex
+    columns = MultiIndex.from_tuples([row.split('_', 1) for row in data.columns])
+    columns.names = ['GEO Accession', 'index']
+    data.columns = columns
     data = data.T
 
     # make sure the same index gets used
diff --git a/README.md b/README.md
index 4df7bbe1..e0e715a9 100644
--- a/README.md
+++ b/README.md
@@ -9,7 +9,7 @@ The Gaussian processes framework in Python.
 * Travis-CI [unit-tests](https://travis-ci.org/SheffieldML/GPy)
 * [![licence](https://img.shields.io/badge/licence-BSD-blue.svg)](http://opensource.org/licenses/BSD-3-Clause)
 
-[![develstat](https://travis-ci.org/SheffieldML/GPy.svg?branch=devel)](https://travis-ci.org/SheffieldML/GPy) [![appveyor](https://ci.appveyor.com/api/projects/status/662o6tha09m2jix3/branch/deploy?svg=true)](https://ci.appveyor.com/project/mzwiessele/gpy/branch/deploy) [![covdevel](http://codecov.io/github/SheffieldML/GPy/coverage.svg?branch=devel)](http://codecov.io/github/SheffieldML/GPy?branch=devel) [![Research software impact](http://depsy.org/api/package/pypi/GPy/badge.svg)](http://depsy.org/package/python/GPy) [![Code Health](https://landscape.io/github/SheffieldML/GPy/devel/landscape.svg?style=flat)](https://landscape.io/github/SheffieldML/GPy/devel)
+[![develstat](https://travis-ci.org/SheffieldML/GPy.svg?branch=devel)](https://travis-ci.org/SheffieldML/GPy) [![appveyor](https://ci.appveyor.com/api/projects/status/662o6tha09m2jix3/branch/deploy?svg=true)](https://ci.appveyor.com/project/mzwiessele/gpy/branch/deploy) [![coverallsdevel](https://coveralls.io/repos/github/SheffieldML/GPy/badge.svg?branch=devel)](https://coveralls.io/github/SheffieldML/GPy?branch=devel) [![covdevel](http://codecov.io/github/SheffieldML/GPy/coverage.svg?branch=devel)](http://codecov.io/github/SheffieldML/GPy?branch=devel) [![Research software impact](http://depsy.org/api/package/pypi/GPy/badge.svg)](http://depsy.org/package/python/GPy) [![Code Health](https://landscape.io/github/SheffieldML/GPy/devel/landscape.svg?style=flat)](https://landscape.io/github/SheffieldML/GPy/devel)
 
 ## Updated Structure
 
diff --git a/appveyor.yml b/appveyor.yml
index 5a0f0577..f7f7b7ab 100644
--- a/appveyor.yml
+++ b/appveyor.yml
@@ -1,6 +1,8 @@
 environment:
   pip_access:
     secure: 8/ZjXFwtd1S7ixd7PJOpptupKKEDhm2da/q3unabJ00=
+  COVERALLS_REPO_TOKEN:
+    secure: d3Luic/ESkGaWnZrvWZTKrzO+xaVwJWaRCEP0F+K/9DQGPSRZsJ/Du5g3s4XF+tS
   gpy_version: 1.2.1
   matrix:
     - PYTHON_VERSION: 2.7
@@ -25,6 +27,7 @@ install:
  - python -m pip install paramz
  - python -m pip install nose-show-skipped
  - python -m pip install coverage
+ - python -m pip install coveralls
  - python -m pip install codecov
  - python -m pip install twine
  - "python setup.py develop"
diff --git a/travis_tests.py b/travis_tests.py
index 3c1c5c95..16713962 100644
--- a/travis_tests.py
+++ b/travis_tests.py
@@ -36,5 +36,5 @@ matplotlib.use('agg')
 import nose, warnings
 with warnings.catch_warnings():
     warnings.simplefilter("ignore")
-    nose.main('GPy', defaultTest='GPy/testing/', argv=['', '--show-skipped'])
+    nose.main('GPy', defaultTest='GPy/testing', argv=['', '--show-skipped'])