Fix selftest errors (#63)

* fix selftests * pass path_save_data as a string to the checkpoint
GrainLearning · Jan 17, 2024 · 610a5b6 · 610a5b6
1 parent 9d342ae
commit 610a5b6
Show file tree

Hide file tree

Showing 3 changed files with 18 additions and 14 deletions.
diff --git a/grainlearning/rnn/train.py b/grainlearning/rnn/train.py
@@ -142,7 +142,7 @@ def train_without_wandb(preprocessor: Preprocessor, config = None, model: tf.ker
             restore_best_weights=True,
         )
     checkpoint = tf.keras.callbacks.ModelCheckpoint(
-                path_save_data,
+                str(path_save_data),
                 monitor='val_loss',
                 save_best_only=True,
                 save_weights_only=config['save_weights_only']

diff --git a/tests/unit/test_iterative_bayesian_filter.py b/tests/unit/test_iterative_bayesian_filter.py
@@ -130,7 +130,7 @@ def test_run_inference():
     #: Create the iterative bayesian filter from a dictionary
     ibf_cls = IterativeBayesianFilter.from_dict(
         {
-            "inference": {"ess_target": 0.5},
+            "inference": {"ess_target": 0.5, "scale_cov_with_max": True},
             "sampling": {"max_num_components": 5},
             "initial_sampling": 'halton',
         }
@@ -151,7 +151,7 @@ def test_run_inference():
     #: Assert that the inference runs correctly if a proposal density is provided
     ibf_cls = IterativeBayesianFilter.from_dict(
         {
-            "inference": {"ess_target": 0.5},
+            "inference": {"ess_target": 0.5, "scale_cov_with_max": True},
             "sampling": {"max_num_components": 5},
             "initial_sampling": 'halton',
             "proposal": np.array([0.5, 0.2, 0.3])

diff --git a/tests/unit/test_sampling.py b/tests/unit/test_sampling.py
@@ -108,7 +108,8 @@ def test_regenerate_params():
     )
 
     #: Initialize a Gaussian Mixture Model object
-    gmm_cls = GaussianMixtureModel(max_num_components=2, covariance_type="full", random_state=100, expand_factor=2)
+    gmm_cls = GaussianMixtureModel(max_num_components=2, covariance_type="tied", init_params="k-means++",
+                                   random_state=100, expand_factor=2)
 
     #: Generate the initial parameter samples
     system_cls.param_data = generate_params_qmc(system_cls, system_cls.num_samples)
@@ -127,10 +128,10 @@ def test_regenerate_params():
         new_params,
         np.array(
             [
-                [2.50061801e+06, 1.92539376e-01],
-                [5.40525882e+06, 3.10537276e-01],
-                [3.46458943e+06, 3.76945456e-01],
-                [5.66254261e+06, 2.67135646e-01]
+                [3.50816173e+06, 2.49928167e-01],
+                [4.87521971e+06, 3.09866017e-01],
+                [2.57761751e+06, 3.26589170e-01],
+                [5.39169338e+06, 2.52627014e-01]
             ]
         ),
         rtol=0.001,
@@ -140,7 +141,8 @@ def test_regenerate_params():
     proposal = np.array([0.0, 0.5, 0.4, 0.1])
 
     #: Initialize again a Gaussian Mixture Model object with slice sampling activated
-    gmm_cls = GaussianMixtureModel(max_num_components=2, covariance_type="full", expand_factor=10, slice_sampling=True)
+    gmm_cls = GaussianMixtureModel(max_num_components=2, covariance_type="tied", init_params="k-means++",
+                                   expand_factor=10, slice_sampling=True)
 
     #: Generate the initial parameter samples
     system_cls.param_data = generate_params_qmc(system_cls, system_cls.num_samples)
@@ -155,8 +157,8 @@ def test_regenerate_params():
             [
                 [5.50000000e+06, 2.93333333e-01],
                 [3.25000000e+06, 3.96666667e-01],
-                [6.90625000e+06, 2.47407407e-01],
-                [4.23437500e+06, 3.35432099e-01]
+                [5.07812500e+06, 3.20123457e-01],
+                [4.55078125e+06, 3.30329218e-01]
             ]
         ),
         rtol=0.001,
@@ -198,9 +200,11 @@ def test_draw_samples_within_bounds():
         new_params,
         np.array(
             [
-                [6.442085e+06, 3.272525e-01],
-                [5.109979e+06, 2.736598e-01],
-                [4.379674e+06, 3.722799e-01]
+                [3.72038687e+06, 2.50311001e-01],
+                [5.31282617e+06, 3.30634332e-01],
+                [6.04092450e+06, 2.61269286e-01],
+                [5.21447138e+06, 2.84730873e-01],
+                [3.60796701e+06, 2.81671695e-01]
             ]
         ),
         rtol=0.001,