to pull req

aimclub · Jul 16, 2024 · 2e2ad91 · 2e2ad91
1 parent 49d0570
commit 2e2ad91
Show file tree

Hide file tree

Showing 5 changed files with 4 additions and 13 deletions.
diff --git a/fedot_ind/core/models/early_tc/base_early_tc.py b/fedot_ind/core/models/early_tc/base_early_tc.py
@@ -85,9 +85,9 @@ def _predict(self, X, training=True):
         estimator_indices, offset = self._select_estimators(X, training)
         if not training:
             self._estimator_for_predict = estimator_indices
-        prediction = zip(
+        prediction = (np.stack(array_list) for array_list in zip(
             *[self._predict_one_slave(X, i, offset) for i in estimator_indices] # check boundary
-        )
+        ))
         return prediction # see the output in _predict_one_slave
 
     def _consecutive_count(self, predicted_labels: List[np.array]):

diff --git a/fedot_ind/core/models/early_tc/ecec.py b/fedot_ind/core/models/early_tc/ecec.py
@@ -23,11 +23,8 @@ def _predict_one_slave(self, X, i, offset=0):
 
     def _predict(self, X, training=False):
         predicted_labels, predicted_probas, reliabilities = super()._predict(X, training)
-        reliabilities = np.stack(reliabilities)
         confidences = 1 - np.cumprod(1 - reliabilities, axis=0)
         non_confident = confidences < self.confidence_thresholds[:len(predicted_labels), None]
-        predicted_labels = np.stack(predicted_labels)
-        predicted_probas = np.stack(predicted_probas)
         return predicted_labels, predicted_probas, non_confident, confidences
 
     def predict_proba(self, X):

diff --git a/fedot_ind/core/models/early_tc/economy_k.py b/fedot_ind/core/models/early_tc/economy_k.py
@@ -78,7 +78,6 @@ def _get_prediction_time(self, X, cluster_centroids, i):
 
     def predict_proba(self, X):
         probas, times, _ = self._predict(X, training=False)
-        probas, times = np.stack(probas), np.stack(times)
         return super().predict_proba(probas, times)
 
     def _transform_score(self, time):

diff --git a/fedot_ind/core/models/early_tc/prob_threshold.py b/fedot_ind/core/models/early_tc/prob_threshold.py
@@ -24,8 +24,6 @@ def predict_proba(self, X):
 
     def _predict(self, X, training=True):
         predicted_probas, predicted_labels = super()._predict(X, training)
-        predicted_probas = np.stack(predicted_probas)
-        predicted_labels = np.stack(predicted_labels)
         non_acceptance = self._consecutive_count(predicted_labels) < self.consecutive_predictions
         double_check = predicted_probas.max(axis=-1) > self.probability_threshold
         non_acceptance[non_acceptance & double_check] = False

diff --git a/fedot_ind/core/models/early_tc/teaser.py b/fedot_ind/core/models/early_tc/teaser.py
@@ -43,13 +43,10 @@ def _form_X_oc(self, predicted_probas):
 
     def _predict(self, X, training=False):
         estimator_indices, offset = self._select_estimators(X)
-        X_ocs, predicted_probas, predicted_labels = zip(
+        X_ocs, predicted_probas, predicted_labels = map(np.stack, zip(
             *[self._predict_one_slave(X, i, offset) for i in estimator_indices]
-        )
+        ))
         non_acceptance = self._consecutive_count(predicted_labels) < self.consecutive_predictions
-        X_ocs = np.stack(X_ocs)
-        predicted_probas = np.stack(predicted_probas)
-        predicted_labels = np.stack(predicted_labels)
         final_verdicts = np.zeros((len(estimator_indices), X.shape[0]))
         # for each point of estimation 
         for i in range(predicted_labels.shape[0]):