Temporarily suppress GaussianProcessRegressionModel in stochastic_pro…

…cess_properties_test.

PiperOrigin-RevId: 281353351

tensorflow_probability/python/distributions/stochastic_process_properties_test.py

@@ -197,6 +197,7 @@ def gaussian_process_regression_models(draw,
       feature_ndims=feature_ndims,
       enable_vars=enable_vars,
       name='index_points'))
+  hp.note('Index points: {}'.format(index_points))
 
   observation_index_points = draw(
       kernel_hps.kernel_input(
@@ -206,6 +207,7 @@ def gaussian_process_regression_models(draw,
           feature_ndims=feature_ndims,
           enable_vars=enable_vars,
           name='observation_index_points'))
+  hp.note('Observation index points: {}'.format(observation_index_points))
 
   observations = draw(kernel_hps.kernel_input(
       batch_shape=compatible_batch_shape,
@@ -218,12 +220,15 @@ def gaussian_process_regression_models(draw,
       feature_ndims=0,
       enable_vars=enable_vars,
       name='observations'))
+  hp.note('Observations: {}'.format(observations))
 
   params = draw(broadcasting_params(
       'GaussianProcessRegressionModel',
       compatible_batch_shape,
       event_dim=event_dim,
       enable_vars=enable_vars))
+  hp.note('Params: {}'.format(params))
+
   gp = tfd.GaussianProcessRegressionModel(
       kernel=k,
       index_points=index_points,
@@ -296,6 +301,10 @@ class StochasticProcessParamsAreVarsTest(test_util.TestCase):
           hp.HealthCheck.filter_too_much,
           hp.HealthCheck.data_too_large])
   def testProcess(self, process_name, data):
+    if process_name == 'GaussianProcessRegressionModel':
+      import unittest  # pylint: disable=g-import-not-at-top
+      raise unittest.case.SkipTest('b/144181034')
+
     if tf.executing_eagerly() != (FLAGS.tf_mode == 'eager'):
       return
     seed = test_util.test_seed()
@@ -332,6 +341,7 @@ def testProcess(self, process_name, data):
           'method `sample` of `{}`'.format(process),
           max_permissible=excessive_usage_count(process_name)):
         sample = process.sample(seed=seed)
+    hp.note('Drew sample {}'.format(sample))
     if process.reparameterization_type == tfd.FULLY_REPARAMETERIZED:
       grads = tape.gradient(sample, process.variables)
       for grad, var in zip(grads, process.variables):