[MNT] Make Multi Rec test in deep clustering faster (#2315)

* Update test_clusterer_features.py

* precommit

* fixed bug

* remove test replace by combination

* re do test

Author: hadifawaz1999

aeon/clustering/deep_learning/_ae_fcn.py

@@ -489,17 +489,17 @@ def _get_test_params(cls, parameter_set="default"):
             Each dict are parameters to construct an "interesting" test instance, i.e.,
             `MyClass(**params)` or `MyClass(**params[i])` creates a valid test instance.
         """
-        param1 = {
-            "n_epochs": 1,
+        param = {
+            "n_epochs": 2,
             "batch_size": 4,
             "use_bias": False,
-            "n_layers": 1,
-            "n_filters": 4,
-            "kernel_size": 2,
+            "n_layers": 2,
+            "n_filters": [2, 2],
+            "kernel_size": [2, 2],
             "padding": "same",
             "strides": 1,
-            "latent_space_dim": 4,
+            "latent_space_dim": 2,
             "estimator": DummyClusterer(n_clusters=2),
         }
 
-        return [param1]
+        return [param]

aeon/clustering/deep_learning/_ae_resnet.py

@@ -502,11 +502,11 @@ def _get_test_params(cls, parameter_set="default"):
             `MyClass(**params)` or `MyClass(**params[i])` creates a valid test instance.
         """
         param = {
-            "n_epochs": 1,
+            "n_epochs": 2,
             "batch_size": 4,
-            "n_residual_blocks": 1,
+            "n_residual_blocks": 2,
             "n_conv_per_residual_block": 1,
-            "n_filters": 1,
+            "n_filters": [2, 2],
             "kernel_size": 2,
             "use_bias": False,
             "estimator": DummyClusterer(n_clusters=2),

aeon/clustering/deep_learning/tests/test_clusterer_features.py renamed to aeon/clustering/deep_learning/tests/test_deep_clusterer_features.py

@@ -14,13 +14,12 @@
 def test_multi_rec_fcn():
     """Tests whether multi-rec loss works fine or not."""
     X = np.random.random((100, 5, 2))
-    clst = AEFCNClusterer(
-        n_clusters=2, n_epochs=10, n_filters=[2, 3, 4], loss="multi_rec"
-    )
+    clst = AEFCNClusterer(**AEFCNClusterer._get_test_params()[0], loss="multi_rec")
     clst.fit(X)
-    assert (
-        clst.history["loss"][0] > clst.history["loss"][9]
-    )  # Check if loss is decreasing.
-    clst = AEResNetClusterer(n_clusters=2, n_epochs=10, loss="multi_rec")
+    assert isinstance(clst.history["loss"][-1], float)
+
+    clst = AEResNetClusterer(
+        **AEResNetClusterer._get_test_params()[0], loss="multi_rec"
+    )
     clst.fit(X)
-    assert clst.history["loss"][0] > clst.history["loss"][9]
+    assert isinstance(clst.history["loss"][-1], float)

aeon/networks/_ae_resnet.py

@@ -225,7 +225,7 @@ def build_network(self, input_shape, **kwargs):
                 conv = tf.keras.layers.Conv1D(
                     filters=self._n_filters[d],
                     kernel_size=self._kernel_size[c],
-                    strides=self._strides[d],
+                    strides=self._strides[c],
                     padding=self._padding[c],
                     dilation_rate=self._dilation_rate[c],
                 )(x)
@@ -290,7 +290,7 @@ def build_network(self, input_shape, **kwargs):
                 conv = tf.keras.layers.Conv1DTranspose(
                     filters=self._n_filters[d],
                     kernel_size=self._kernel_size[c],
-                    strides=self._strides[d],
+                    strides=self._strides[c],
                     padding=self._padding[c],
                     dilation_rate=self._dilation_rate[c],
                 )(x)

aeon/networks/_resnet.py

@@ -209,7 +209,7 @@ def build_network(self, input_shape, **kwargs):
                 conv = tf.keras.layers.Conv1D(
                     filters=self._n_filters[d],
                     kernel_size=self._kernel_size[c],
-                    strides=self._strides[d],
+                    strides=self._strides[c],
                     padding=self._padding[c],
                     dilation_rate=self._dilation_rate[c],
                 )(x)